From 80ddc3541251c9bc0bc37f57029f9fa4c63d5ed0 Mon Sep 17 00:00:00 2001 From: Jeroen Schweitzer Date: Fri, 6 Mar 2026 20:56:39 +0100 Subject: [PATCH] docs(workshops): add Where's the Fun workshop outputs 5 rounds, 9 agents + Qatux + SI, 24 decisions locked. Full round transcripts and workshop outcomes summary. Co-Authored-By: Claude Opus 4.6 --- .../interview-supplement-round4.md | 70 +++ .../wheres-the-fun/interview-supplement.md | 70 +++ .../wheres-the-fun/lead-interview.md | 199 +++++++ .../workshops/wheres-the-fun/round-1-notes.md | 160 +++++ .../workshops/wheres-the-fun/round-3-notes.md | 363 ++++++++++++ .../workshops/wheres-the-fun/round-4-notes.md | 400 +++++++++++++ .../workshops/wheres-the-fun/round-5-notes.md | 413 +++++++++++++ .../wheres-the-fun/round1-araminta.md | 59 ++ .../wheres-the-fun/round1-gestalt.md | 54 ++ docs/workshops/wheres-the-fun/round1-gore.md | 57 ++ .../wheres-the-fun/round1-mellanie.md | 57 ++ docs/workshops/wheres-the-fun/round1-miri.md | 43 ++ docs/workshops/wheres-the-fun/round1-nigel.md | 47 ++ docs/workshops/wheres-the-fun/round1-ozzie.md | 58 ++ docs/workshops/wheres-the-fun/round1-paula.md | 60 ++ docs/workshops/wheres-the-fun/round1-tyre.md | 72 +++ .../wheres-the-fun/round3-araminta.md | 125 ++++ .../wheres-the-fun/round3-gestalt.md | 192 ++++++ docs/workshops/wheres-the-fun/round3-gore.md | 97 ++++ .../wheres-the-fun/round3-mellanie.md | 138 +++++ docs/workshops/wheres-the-fun/round3-miri.md | 151 +++++ docs/workshops/wheres-the-fun/round3-nigel.md | 146 +++++ docs/workshops/wheres-the-fun/round3-ozzie.md | 220 +++++++ docs/workshops/wheres-the-fun/round3-paula.md | 184 ++++++ docs/workshops/wheres-the-fun/round3-tyre.md | 246 ++++++++ .../wheres-the-fun/round4-araminta.md | 85 +++ .../wheres-the-fun/round4-gestalt.md | 106 ++++ docs/workshops/wheres-the-fun/round4-gore.md | 87 +++ .../wheres-the-fun/round4-interview.md | 106 ++++ .../wheres-the-fun/round4-mellanie.md | 118 ++++ docs/workshops/wheres-the-fun/round4-miri.md | 106 ++++ docs/workshops/wheres-the-fun/round4-nigel.md | 117 ++++ docs/workshops/wheres-the-fun/round4-ozzie.md | 121 ++++ docs/workshops/wheres-the-fun/round4-paula.md | 108 ++++ docs/workshops/wheres-the-fun/round4-tyre.md | 175 ++++++ .../wheres-the-fun/round5-araminta.md | 467 +++++++++++++++ .../wheres-the-fun/round5-gestalt.md | 492 ++++++++++++++++ docs/workshops/wheres-the-fun/round5-gore.md | 363 ++++++++++++ .../wheres-the-fun/round5-interview.md | 102 ++++ .../wheres-the-fun/round5-mellanie.md | 494 ++++++++++++++++ docs/workshops/wheres-the-fun/round5-miri.md | 483 +++++++++++++++ docs/workshops/wheres-the-fun/round5-nigel.md | 406 +++++++++++++ docs/workshops/wheres-the-fun/round5-ozzie.md | 417 +++++++++++++ docs/workshops/wheres-the-fun/round5-paula.md | 421 ++++++++++++++ docs/workshops/wheres-the-fun/round5-tyre.md | 548 ++++++++++++++++++ .../wheres-the-fun-workshop-brief.md | 198 +++++++ .../wheres-the-fun/workshop-outcomes.md | 208 +++++++ 47 files changed, 9409 insertions(+) create mode 100644 docs/workshops/wheres-the-fun/interview-supplement-round4.md create mode 100644 docs/workshops/wheres-the-fun/interview-supplement.md create mode 100644 docs/workshops/wheres-the-fun/lead-interview.md create mode 100644 docs/workshops/wheres-the-fun/round-1-notes.md create mode 100644 docs/workshops/wheres-the-fun/round-3-notes.md create mode 100644 docs/workshops/wheres-the-fun/round-4-notes.md create mode 100644 docs/workshops/wheres-the-fun/round-5-notes.md create mode 100644 docs/workshops/wheres-the-fun/round1-araminta.md create mode 100644 docs/workshops/wheres-the-fun/round1-gestalt.md create mode 100644 docs/workshops/wheres-the-fun/round1-gore.md create mode 100644 docs/workshops/wheres-the-fun/round1-mellanie.md create mode 100644 docs/workshops/wheres-the-fun/round1-miri.md create mode 100644 docs/workshops/wheres-the-fun/round1-nigel.md create mode 100644 docs/workshops/wheres-the-fun/round1-ozzie.md create mode 100644 docs/workshops/wheres-the-fun/round1-paula.md create mode 100644 docs/workshops/wheres-the-fun/round1-tyre.md create mode 100644 docs/workshops/wheres-the-fun/round3-araminta.md create mode 100644 docs/workshops/wheres-the-fun/round3-gestalt.md create mode 100644 docs/workshops/wheres-the-fun/round3-gore.md create mode 100644 docs/workshops/wheres-the-fun/round3-mellanie.md create mode 100644 docs/workshops/wheres-the-fun/round3-miri.md create mode 100644 docs/workshops/wheres-the-fun/round3-nigel.md create mode 100644 docs/workshops/wheres-the-fun/round3-ozzie.md create mode 100644 docs/workshops/wheres-the-fun/round3-paula.md create mode 100644 docs/workshops/wheres-the-fun/round3-tyre.md create mode 100644 docs/workshops/wheres-the-fun/round4-araminta.md create mode 100644 docs/workshops/wheres-the-fun/round4-gestalt.md create mode 100644 docs/workshops/wheres-the-fun/round4-gore.md create mode 100644 docs/workshops/wheres-the-fun/round4-interview.md create mode 100644 docs/workshops/wheres-the-fun/round4-mellanie.md create mode 100644 docs/workshops/wheres-the-fun/round4-miri.md create mode 100644 docs/workshops/wheres-the-fun/round4-nigel.md create mode 100644 docs/workshops/wheres-the-fun/round4-ozzie.md create mode 100644 docs/workshops/wheres-the-fun/round4-paula.md create mode 100644 docs/workshops/wheres-the-fun/round4-tyre.md create mode 100644 docs/workshops/wheres-the-fun/round5-araminta.md create mode 100644 docs/workshops/wheres-the-fun/round5-gestalt.md create mode 100644 docs/workshops/wheres-the-fun/round5-gore.md create mode 100644 docs/workshops/wheres-the-fun/round5-interview.md create mode 100644 docs/workshops/wheres-the-fun/round5-mellanie.md create mode 100644 docs/workshops/wheres-the-fun/round5-miri.md create mode 100644 docs/workshops/wheres-the-fun/round5-nigel.md create mode 100644 docs/workshops/wheres-the-fun/round5-ozzie.md create mode 100644 docs/workshops/wheres-the-fun/round5-paula.md create mode 100644 docs/workshops/wheres-the-fun/round5-tyre.md create mode 100644 docs/workshops/wheres-the-fun/wheres-the-fun-workshop-brief.md create mode 100644 docs/workshops/wheres-the-fun/workshop-outcomes.md diff --git a/docs/workshops/wheres-the-fun/interview-supplement-round4.md b/docs/workshops/wheres-the-fun/interview-supplement-round4.md new file mode 100644 index 000000000..cfe564bf7 --- /dev/null +++ b/docs/workshops/wheres-the-fun/interview-supplement-round4.md @@ -0,0 +1,70 @@ +# Interview Supplement — Round 4 Decisions +## Where's the Fun? Workshop | 2026-03-05 + +The full Round 4 interview transcript is at docs/workshops/wheres-the-fun/round4-interview.md. This supplement highlights the 15 decisions locked and their implications. + +--- + +## The 15 Decisions + +### Proof-of-Life Scope + +1. **Proof-of-life = generator + graphics, not hand-built slice.** The v0.1 lesson: descoping led to the wrong game. v0.2 proves the foundation (auto-generated locations at scale + legible characters) first, then builds the game on top. Similar reasoning to v0.1's narrow scope is explicitly rejected. + +2. **Skills + bookmark only for character creation.** Family/culture/religion deferred. + +3. **Religion is NOT a game system.** It was a CK3 reference point, not a design requirement. Remove from design scope. + +4. **Tycoon is the v0.2 bookmark.** Zero investigation content. Clean break from detective/smuggler. Tycoon naturally blends Active (manage business), WFH (remote investments via insert), and Gig (one-off deals). + +### Skills & Voice + +5. **Skills affect outcome (mostly C).** Everyone sees the same verbs. Skills determine how well you do. Some advanced verbs may still be gated — spec needed for which ones. + +6. **Voice: culture-driven, job modifies.** The character IS their background. Job adds a layer. Voice cards authored at the culture level with job-specific modifiers. A Krenn tycoon sounds like a Krenn person who runs businesses, not a generic tycoon. NOTE: this is inverted from "job sets base, culture modifies" — culture is primary. + +### Content Architecture + +7. **ALL NPCs are generated. No named characters.** Kael doesn't exist. The generator produces NPCs that fit positions based on location characteristics. A smuggling operation exists because the geography enables it; NPCs fill roles organically. + +8. **Generative AI for NPC content templating.** Culture vectors, tone, accents as templating dimensions. Limited vocabulary acceptable at first. The copy pool will be enormous but AI-assisted. + +9. **Possible in-game ollama for live NPC dialogue.** Deferred but the door is explicitly open. A dressed-down LLM running in-game for dynamic dialogue is on the table for later investigation. + +10. **Quietly responsive world, not indifferent.** Gore's Kenshi-indifference premise rejected. The world doesn't care globally but notices locally. Primary social contacts (colleagues, neighbors) develop responsiveness over time. Gradient of caring based on social proximity. Even in early builds with test users, the world is never truly indifferent. + +### Visual & Identity + +11. **Full character customization.** Hair, clothing, colors. The creation screen is part of identity investment. Readability at tile scale solved through outline/highlight, not by limiting customization. + +12. **Setting delivery: both layers.** World shows it through visuals and behavior; insert names and contextualizes. Araminta (visual) and Mellanie (insert copy) work in parallel. + +13. **First Settled Reach moment: apartment + insert activation.** The apartment is auto-generated (reflects wealth, location, lifestyle). Then the insert powers on. Two intimate, personal moments before you leave the room. + +14. **Groundhog Day alarm clock homage.** *Click* pa-pa pa-pa, cut short. First game day only. "New day, new start, new chances" with a wink. + +### The Rimworld Model + +15. **Player choices ARE the content.** The question "is Phase 1 dev-sequencing or player experience?" is wrong-framed. The world is full of opportunity; the player's choices are the content. Rimworld model: one authored starting beat (alarm clock), then agency and options. A job is rails to take off from, not a script to follow. + +--- + +## What These Decisions Mean For Each Domain + +**Systems (Gestalt):** Skills-verb coupling is outcome-based. VerbPriorityProfile still matters (job-aware ordering) but all verbs visible to all players. The verb system is more learnable and less gating. Tycoon bookmark means designing economic verbs first. + +**Player Experience (Ozzie):** Wow moments must work for a tycoon, not a detective. "First Day" is apartment + insert activation + arriving at your first business appointment. The Consequence moment is economic (a deal went wrong, prices shifted, someone remembers what you did). + +**Narrative (Paula):** No named NPCs changes everything. FRIEND pattern becomes a generator template, not authored character. Phase Zero warmth must emerge from generated NPCs with limited vocabulary. The moral arc still works but is triggered by emergent relationships, not authored relationships. + +**Themes (Gore):** Quietly responsive, not indifferent. Consequence as theme survives. The gradient of caring (world → district → neighbors → colleagues → friends) is the emotional architecture. Transhumanist hooks via skill_ceiling concept. + +**Replayability (Nigel):** Tycoon-only bookmark for v0.2 limits cross-career testing. But generator-first approach means every seed produces a different world — replayability comes from world variation, not career variation, in v0.2. + +**Worldbuilding (Miri):** Zone identity spec is urgent — the generator needs it before it can produce "the Settled Reach." Both-layer setting delivery means visual + insert work in parallel. Auto-generated apartments add a new authoring dimension (what does wealth look like in different cultures?). + +**Architecture (Tyre):** Generator is the v0.2 proof-of-life, not a hand-built slice. This is more work upfront but proves the foundation. NPC generation pipeline + character creation (skills + bookmark) + one diegetic tool suite for tycoon. + +**Visual (Araminta):** Full character customization + NPC legibility + auto-generated apartments. Three visual production challenges simultaneously. Readability at tile scale with custom characters is the hardest unsolved problem. + +**Copy (Mellanie):** Culture-driven voice with job modifier. No named characters means template-based content. AI-assisted templating pipeline. Insert copy for setting delivery in parallel with visual. Limited NPC vocabulary acceptable at first. diff --git a/docs/workshops/wheres-the-fun/interview-supplement.md b/docs/workshops/wheres-the-fun/interview-supplement.md new file mode 100644 index 000000000..014c7ef1b --- /dev/null +++ b/docs/workshops/wheres-the-fun/interview-supplement.md @@ -0,0 +1,70 @@ +# Interview Supplement — Post-Round 3 Direction +## Where's the Fun? Workshop | 2026-03-05 + +After Round 3 proposals were submitted, the facilitator resolved the 4 open tensions with Jeroen and collected additional design direction. This document captures all of that verbatim. + +--- + +## Tension 1: Career Bookmarks — Resolved as "Different Framing Needed" + +The wiki-review workshop (Feb 12) established 8 fluid PC archetypes. Jeroen rejected the archetype list as the right framing for bookmarks. + +**Jeroen's direction:** "CK3 is a lot more apt here than even just the bookmarks. It allows for family creation, religion selection, culture selection, skill selection, skill level setting. We don't want D&D style base stats, but shooting proficiency will help when handling a gun. Somewhere between CK3 and The Sims I feel is the sweet spot with skills. This is even more important than job/archetype. The player plays how they want. The job/archetype/starting environment determine what is in the toolbox, not what the character is like. Borrow a character creation budget from CK3 (with console cheat options) to bound people from becoming OP." + +**Key principles:** +- Character creation (skills, culture, family, budget) is MORE fundamental than job/bookmark choice +- Skills are proficiencies (shooting, hacking, social manipulation, mechanical repair), NOT D&D base stats +- Job/archetype/starting environment = the toolbox (inserts, contacts, access), NOT the character's personality +- CK3-style point budget with console cheat option for unconstrained builds +- The archetype list from the wiki workshop is not exhaustive and needs revisiting for the life-sim pivot + +--- + +## Tension 2: First 30 Minutes — Resolved as "Groundhog Day" + +**Jeroen's direction:** "Groundhog Day. The day starts with an alarm clock, and a calendar ping that says you are expected at onboarding location X (bank/job/train station) or nowhere if you go clean, and you then have to open a job board or phone book to start figuring it out through your insert (experienced player warning on this mode)." + +**The flow:** +1. Character creation (CK3-style) +2. Pick starting situation (bookmark = job + location + relationships + tools) +3. Day 1: Alarm clock. Calendar ping with destination (or nothing for clean start). +4. Bookmarked: go to your appointment, onboarding begins +5. Clean start: no appointment, figure it out via job board/phone book/insert. Experienced player warning. + +--- + +## Tension 3: Missions — Resolved as "Emergent Life, Not Missions" + +**Jeroen's direction:** "Let me rephrase mission with emergent things that happen in the context of what the player chooses to do. Sometimes this will be a mission (like an assassination target) most of the times it will be more of a 'I have a plan, let's see if it flies' situation. The first situation is a mission with a reward layer, the second is, you spent a day cooking at the bar you work at, met some people, ended a fight in an alley and then went home with finally enough money saved to buy that car the day after." + +**Three career models coexisting:** +1. **Active** — you're at the workplace, the work IS the gameplay (cooking at the bar, patrolling as law enforcement) +2. **WFH/Remote** — portable work you can do anywhere, income while exploring (hacking contracts, writing, remote consulting via insert) +3. **Gig/Freelance** — pick jobs from a board, outcomes vary by quality, no fixed hours (smuggling runs, fixer contracts, bounty hunting) + +**Reference games confirmed by Jeroen:** +- The Sims (life sim framing, relationship systems, career models) +- CK3 (character creation, bookmarks, dynasty, skill budgets) +- Rimworld (storyteller escalation, dramatic timing, AI-driven events) +- Dwarf Fortress (emergent history, world doesn't care about you) +- Kenshi (sandbox survival, the world is indifferent, you are not special) + +--- + +## Tension 4: Endgame — Resolved as "Include Hooks" + +Plant small seeds in v0.2 (skill progression hints, insert upgrade paths) that point toward the transhumanist ladder without implementing it. Don't foreclose the path; don't build it yet. + +--- + +## Critical Sequencing — The Two-Phase Approach + +**Jeroen's explicit sequencing for v0.2:** + +**Phase 1: The uncaring world.** The generator runs. Geography -> infrastructure -> zones -> population -> routines. NPCs go about their days. The economy ticks. The simulation doesn't know or care that a player exists. This must feel natural-but-artificial before anything authored enters. Primitive components tied together producing emergent complexity. Reference: the Generator Architecture workshop established the Cities Skylines top-down pipeline model with 14 confirmed D-records. + +**Phase 2: Authored content injected.** Once the world runs and feels alive, quest/mission components are randomized into the simulation — not as scripted sequences but as authored ingredients (triangle templates, FRIEND arcs, contradiction arcs) placed by the generator into the living world. The Rimworld storyteller decides WHEN and HOW MUCH PRESSURE, not WHAT HAPPENS. + +**Jeroen's words:** "Emergent complexity like this will start with the uncaring world generating properly. And a lot just happening from primitive components tied together. Once that runs and feels natural (but artificial) is when we start with authored quest and mission components we randomize into the gameplay." + +This is the order. World first, then content. diff --git a/docs/workshops/wheres-the-fun/lead-interview.md b/docs/workshops/wheres-the-fun/lead-interview.md new file mode 100644 index 000000000..2c63815b3 --- /dev/null +++ b/docs/workshops/wheres-the-fun/lead-interview.md @@ -0,0 +1,199 @@ +# Round 2: Interview Transcript +## Where's the Fun? Workshop | 2026-03-05 + +**Interviewer:** Workshop facilitator (consolidated questions from 9 agents) +**Interviewee:** Jeroen (designer, first playtester of v0.1) +**Format:** Interactive interview, questions presented in thematic groups + +--- + +## Group 1: The Vision and the Wall + +### Q1 (Ozzie): Close your eyes. The game is finished and fun. What are you DOING? + +**Jeroen:** None of the offered frames (watching/realizing, feeling implicated, social manipulation) fit. "I want to live in the world and have impact. Own businesses, build relationships, hunt enemies, defend against enemies, build property, get a job (which could be detective). My ideal game would be playing The Sims Rimworld-style, from a first-person perspective (well third, but single character)." + +**Key finding:** The vision is a life sim from a single-character perspective. Detective is a job you might have, not the core loop. The v0.1 vertical slice built around detective/smuggler was a narrowing that lost the broader vision. + +--- + +### Q2 (Gore): What were you actually DOING moment-to-moment in the playtest? + +**Jeroen:** Mix of all three offered options — cycling between aimless walking, fighting the UI, and trying to process information. + +**Key finding:** The actual playtest verb was "navigate confusion." None of the intended verbs (observe, investigate, engage) were operating. + +--- + +### Q3 (Gestalt): At minute 2, what did you think you were supposed to do? + +**Jeroen:** Faint signal — there was something there but too weak to act on. + +**Key finding:** The systems weren't completely silent. Something was reaching the player, but below the threshold of actionability. This suggests the feedback channels exist but are severely underpowered. + +--- + +## Group 2: The Confusion Type + +### Q4 (Gestalt): Was it "I know something is here" or "I have no idea what I'm looking at"? + +**Jeroen:** "Mostly B with a hint of A, layered with the looming realization that the narrow scoping for v0.1 was actually a bad frame to work towards since we lost track of the broader brief." + +**Key finding:** Predominantly uniform opacity (bad confusion), not engaged mystery (good confusion). But the deeper realization was meta-level: the playtest wasn't just confusing in the moment — it revealed that the v0.1 framing itself was wrong. The narrow detective scope had lost track of the life-sim vision. + +--- + +### Q5 (Araminta): Were you unable to READ the screen, or unable to ACT on what you read? + +**Jeroen:** Both equally. + +**Key finding:** Both attention management (visual hierarchy) and agency (what to do) were failing simultaneously. Araminta and the systems designers both have work to do, but visual hierarchy is needed regardless of the design reframe. + +--- + +## Group 3: The Emotional Loop + +### Q6 (Paula): Did Phase 1 feel like a stable world you belonged to? + +**Jeroen:** "I had no idea what dot was who, or to be honest that they were people at all (had I not been involved in the sprints running up to this one)." + +**Key finding:** The NPCs did not register as human beings. They were dots. The entire emotional stack (attachment -> moral arc -> complicity -> dual-lens reveal) has no foundation because the entities on screen are not legible as people. This is a more fundamental failure than anyone anticipated — not "the emotional loop didn't fire" but "the preconditions for recognizing there ARE characters didn't exist." + +--- + +### Q7 (Paula): Did any NPC feel like a person with stakes? + +**Jeroen:** No one. No NPC registered as a person. They were all moving tiles. + +**Key finding:** Zero NPC attachment. The moral arc literally cannot start. The Phase 1-to-2 gates (observing Naia's stress, Maret's anxiety) require the player to be watching specific people closely — but the people aren't legible as people. + +--- + +### Q8 (Gore): Did the complicity threshold ever fire? + +**Jeroen:** "Seen the previous answer that the dots did not register as anything but dots, this question assumes way too much engagement to even start speaking in those terms." + +**Key finding:** The question was premature. Complicity requires recognizing characters as people, recognizing your relationship to them, and then feeling implicated. The playtest never reached step one. The complicity theme exists only in design documents, not in the player experience. This is not fixable with better monologue timing or content — it requires the characters to become visible and legible first. + +--- + +## Group 4: The Monologue System + +### Q9 (Mellanie): Could you connect monologue lines to their triggers? + +**Jeroen:** "It felt like it wanted me to know something, but no idea what it was connected to, what triggered it, and even how to relate the messages to each other since they were coming in scattered." + +**Key finding:** The monologue system was perceived as intentional communication (it wanted to tell the player something) but failed on three axes: no spatial anchor to trigger source, no causal legibility, and no inter-message continuity. The lines couldn't build on each other because there was no thread connecting them. + +--- + +### Q10 (Mellanie): Did the monologue ever feel like your character's thoughts? + +**Jeroen:** Hard to tell — given that the dots weren't people and the world wasn't legible, the question isn't answerable from this playtest. + +**Key finding:** The monologue-as-interiority question is unanswerable until the floor exists. If the world isn't a place and the NPCs aren't people, the character's thoughts about them can't register as interiority. The system needs to be re-evaluated after the fundamental legibility problems are solved. + +--- + +## Group 5: Systems & Architecture + +### Q11 (Tyre): Was monologue the only planned channel for the knowledge graph? + +**Jeroen:** "Always planned for more, like the mystery board with the wires connecting topics, a glossary of characters, a journal tracking communications, tele-communications, insert icons, AR type overlays on the world. The monologue was supposed to be a reasoning nudge and summary tool in the later stages. We descoped for v0.1, which brings me back to that v0.1 was not framed correctly. We learn this and do better for v0.2, possibly with a whole new game loop." + +**Key finding:** The full information architecture was always planned — mystery board, character glossary, journal, comms, insert icons, AR overlays. Monologue was never supposed to carry the feedback burden alone; it was meant as a reasoning nudge in the later stages. The v0.1 descoping stripped the player of all their planned tools and then asked them to play. This confirms the scoping error: v0.1 didn't just narrow the story — it removed the player's interface to the world. + +**Process note:** Jeroen explicitly requested that each Round 3 agent receive the full verbatim transcript of this interview, not just summaries. "I want all finesse communicated." + +--- + +### Q12 (Tyre): Where on the objectives spectrum? Is "no objectives" a principle or scope decision? + +**Jeroen:** "I think the underlying reasoning was that, for replayability, the information systems needed to be framed in-world. This way different jobs, inserts, character budgets would have different tools (that they could aspire and grow towards) and not a GTA-style line leading to objective and large box that checks off mission goals. In fact missions should have varying levels of success. Or even failure declared and sold as success. There would be consequences (enemies, lower pay, no pay, getting fired, an innocent killed or in jail) but it would be possible." + +**Key finding:** "No objectives" was never a design principle against feedback. It was a commitment to diegetic, in-world information tools that vary by job/insert/character build. Different characters would have different HUDs essentially — a detective's neural insert surfaces different information than a smuggler's street contacts network. The absence of objectives in v0.1 was the absence of these diegetic tools, not an intentional design choice for silence. Additionally: missions should have spectrums of success/failure with consequences, not binary pass/fail. A mission can "fail" and the player continues with consequences. This is a life-sim philosophy, not a mission-game philosophy. + +--- + +## Group 6: Verb System & World Legibility + +### Q13 (Gestalt): Did the verb priority fight the detective loop? + +**Jeroen:** Didn't get that far — never reached the point of trying to observe specific NPCs. The dots weren't legible enough as people to target. + +**Key finding:** The verb priority spec contradiction (Talk > Observe at close range) is a real design issue, but it's irrelevant to the current playtest because the player never reached the interaction layer. You can't have a verb problem when the entities aren't recognizable as interactable. + +--- + +### Q14 (Miri): What kind of place did Sova Transit feel like? + +**Jeroen:** Game level with dots. No sense of place at all. + +**Key finding:** Zero environmental storytelling landed. The setting design exists in documents but not in the player experience. Sova Transit's social geography (class dynamics, functional clusters, faction pressures) is invisible. The player saw a tile grid with moving sprites. + +--- + +### Q15 (Nigel): Should the storyteller detect first playthrough and pick a higher-pressure seed? + +**Jeroen:** "Different approach needed. In fact, falling into a mission may be the wrong place to start, since so much of the character will need to evolve from where they start. A detective playthrough should basically be: player starts a game with a shortcut bookmark of starting as law enforcement (or any of the other ones. I feel we need to stop framing everything into the detective-smuggler tension, since that brought us knee deep in the wrong place) but anyways, as an example: after picking that bookmark, the game should start with onboarding at their new job, the enabling one by one of their new inserts, their weapon and weapon qualification training. Tycoon would start with a ping that their bank wants to talk at which point they are handed an inheritance (starting capital) and then the tools to start making investments and buy assets." + +**Key finding:** This is the biggest reframe of the interview. The storyteller seed question is moot because the entire starting premise is changing. Instead of "which variant of the detective mission should you start with," the answer is "you shouldn't start in a mission at all." The game should start with CK3-style bookmarks: choose a career path (law enforcement, tycoon, smuggler, etc.), each with its own onboarding arc that teaches the tools organically. Detective is just one bookmark. The detective-smuggler tension is a storyline within the world, not the frame for the entire game. + +--- + +## Group 7: Remaining Questions + +### Q16 (Ozzie): Was the frustration "not built yet" or "wouldn't have cared even if it was"? + +**Jeroen:** "C... with everything, it felt like the wrong game. I do really want the storylines of the gate builders and the tension between smugglers and the law be in the game and a lot of the patterns would still fit, but the focus felt myopic." + +**Key finding:** Neither "not built yet" nor "wouldn't have cared" — it's "wrong game." The existing storylines (gate builders, smuggler/law tension) should survive as content within a broader world, but the v0.1 framing that made them the sole focus was myopic. The patterns (knowledge graph, perception system, monologue, moral arcs) fit into the broader life-sim vision — they just can't be the entire experience. + +--- + +### Q17 (Araminta): Would a 3-tier visual language have given you a foothold? + +**Jeroen:** Would have helped. + +**Key finding:** Even with all the fundamental problems, visual hierarchy matters. Araminta's argument holds: you can't evaluate any loop until the player can read the screen. Visual hierarchy is needed regardless of the design reframe. This is one thing that carries forward directly into v0.2. + +--- + +### Q18 (Tyre): Pull verb display forward to v0.1? + +**Jeroen:** "This feels like a patch with the wrong priority. I think this and a lot more is needed." + +**Key finding:** Small patches to v0.1 are the wrong frame. The scope of change needed is larger than verb display or monologue tuning. v0.2 needs to be designed from the new vision, and individual v0.1 fixes should be evaluated in that context. + +--- + +## Interview Summary: The Five Revelations + +1. **The vision is a life sim, not a detective game.** Single-character Sims-meets-Rimworld. Detective is one job among many (tycoon, smuggler, law enforcement). The v0.1 detective/smuggler framing was a scope narrowing that lost the broader vision. + +2. **v0.1 was framed wrong.** The descoping didn't just narrow the story — it stripped the player of all their planned information tools (mystery board, journal, AR overlays, comms) and then asked them to play with monologue alone. The playtest failure was predictable in retrospect. + +3. **The world isn't legible.** NPCs are dots, not people. Sova Transit is a game level, not a place. No emotional loop can run on this foundation. The character legibility problem is more fundamental than anyone anticipated. + +4. **"No objectives" was never the principle — diegetic tools were.** Different jobs/inserts provide different information interfaces. The absence of feedback in v0.1 was the absence of these tools, not design purity. + +5. **The game should start with job onboarding, not a mission.** CK3-style bookmarks into career paths, each with onboarding that teaches tools organically. This replaces the storyteller seed variant system as the first-run structure. + +### What Survives + +- The simulation engine (routines, perception, knowledge graph) +- The storylines (gate builders, smuggler/law tension) as world content +- The moral arc patterns (but as consequences within a broader life, not the sole focus) +- The monologue system (as a reasoning nudge, not the primary feedback channel) +- The visual hierarchy work (needed regardless) +- The diegetic information tool vision (mystery board, journal, AR overlays) +- The asymmetric information mechanic (different characters know different things) + +### What Changes + +- The core loop: from "detective puzzle" to "live in the world, have impact" +- The starting experience: from "mission variant" to "job onboarding bookmark" +- The feedback architecture: from "monologue only" to "full diegetic tool suite" +- The vertical slice scope: from "one detective case" to "one career arc with world interaction" +- NPC presentation: dots must become people (art, names, behavior legibility) +- Place presentation: tile grid must become an inhabited space diff --git a/docs/workshops/wheres-the-fun/round-1-notes.md b/docs/workshops/wheres-the-fun/round-1-notes.md new file mode 100644 index 000000000..0703a3d44 --- /dev/null +++ b/docs/workshops/wheres-the-fun/round-1-notes.md @@ -0,0 +1,160 @@ +# Round 1 Notes — Where's the Fun? Workshop +## QATUX — Documenter + +**Workshop:** Where's the Fun? v0.1 Playtest Reckoning +**Round:** 1 — Diagnosis (agents prepare interview questions) +**Date:** 2026-03-05 +**Source files:** `round1-gestalt.md`, `round1-ozzie.md`, `round1-paula.md`, `round1-gore.md`, `round1-nigel.md`, `round1-miri.md`, `round1-tyre.md`, `round1-araminta.md`, `round1-mellanie.md` + +--- + +## Summary + +All 9 agents independently converged on the same diagnosis from their respective domains: **the game has no floor before it asks for a ceiling**. The emotional, relational, and mechanical preconditions for the designed experience don't exist yet. Every agent framed this differently, but they're pointing at the same absence. + +A secondary convergence: **the content pipeline is near-empty**, and the playtest ran on structural scaffolding without authored content. These are related but separable problems. Several agents are careful to distinguish them. + +--- + +## Key Themes + +### 1. The Precondition Problem + +The most consistent theme across all 9 agents: the game's designed payoffs (moral arc, complicity, wow moments, dual-lens reveal, replayability) each require preconditions that are not currently delivered. + +- Paula: the smuggler's moral arc is "temporally displaced" — designed for a player already attached to Kael and the operation. That player never exists in v0.1. The arc needs a relationship establishment phase *before* it can begin. +- Gore: complicity requires a threshold — a before-and-after, a moment you were inside the situation before deciding to be. The playtest may have produced no threshold at all. +- Ozzie: the 6 wow moments are minute-5-to-25 payoffs. The testing wall was minute 8. The wow moments have a floor problem, not a ceiling problem. +- Nigel: replayability is a second-playthrough architecture built on top of first-playthrough investment. If minute 1-8 produces zero attachment, replaying produces "varied nothing." +- Miri: the dual-lens reveal (same space, inverted relationship) requires the first playthrough to have built attachment to Sova Transit as a *real place*. The playtest suggests the player experienced a map, not a place. + +### 2. The Content Pipeline Gap + +Ozzie identified this most directly: 1 of 23 wow-moment deliverables is ready. No authored monologue lines. No FRIEND content. No anomaly detection. The playtest ran on structural systems without the authored layer that makes those systems legible. + +This is distinct from the precondition problem, but intersects with it: even if the arc, the monologue system, and the observation loop are correctly designed, they cannot function without authored content to run through them. + +Mellanie notes the fix order matters: if the delivery system is broken, better content won't save it. But the current playtest can't tell us which is true until lines written to spec actually fire. + +### 3. No Feedback Legibility + +Multiple agents identified the monologue system as failing at its stated purpose — bridging the top-down camera to the character's subjective experience. The failure has at least two components: + +- **Spatial anchoring** (Mellanie, Araminta): monologue fires without a visual anchor to the trigger. The player can't connect the thought to what caused it. +- **Contextless content** (Mellanie, Gestalt): lines fire without establishing *why this thought, now*. Generic lines occupy the moment without earning it. +- **Signal-to-noise overload** (Araminta, Mellanie): too many simultaneous feedback channels of equal visual weight, none of which communicate priority. + +### 4. "No Objectives" — Principle vs Scope + +Tyre surfaced this as an explicit architectural question. The "no objectives, pure observation" philosophy produces paralysis, not discovery. But *why* the philosophy was implemented matters: + +- If it's a **design principle** ("this game is about discovering your own purpose"), the fix is better monologue, spatial feedback, and environmental storytelling. +- If it's a **scope decision** ("we didn't build it yet"), a Tier 2 "threads" system — surfacing the knowledge graph's active contradictions as diegetic character observations — is moderate effort and already architecturally supported. + +No other agent explicitly framed this distinction, but Gore and Paula's questions imply that the first-beat design problem exists regardless of the objectives question. + +### 5. The Verb System May Fight the Detective Loop + +Gestalt identified a specific mechanical contradiction: the verb priority spec puts Talk above Examine NPC at close range. The stated detective loop is "Observe first, Talk later." These directly contradict each other. At close range, `[E]` offers Talk; the player has to back up to mid range to get Observe — but nothing communicates this. The detective's primary verb is hidden behind the social verb. + +--- + +## Recurring Concerns (raised by 3+ agents) + +| Concern | Agents | +|---------|--------| +| No emotional investment in NPCs as people | Paula, Gore, Nigel, Ozzie | +| Player has no purpose, progress, or priority signal | Gestalt, Gore, Araminta, Nigel | +| Moral arc / complicity can't run without preconditions | Paula, Gore, Ozzie | +| Monologue as noise rather than signal | Mellanie, Gestalt, Araminta | +| Spatial anchoring failure (actions, monologue, feedback) | Mellanie, Araminta, Gestalt | +| First 8 minutes is the critical failure zone | All agents | + +--- + +## Unique Insights (domain-specific, not raised by others) + +| Insight | Agent | Significance | +|---------|-------|-------------| +| Verb priority spec directly contradicts detective loop | Gestalt | Specific mechanic — Talk overrides Observe at close range. May be a design decision or an error. | +| Variant A (quiet morning) is a second-playthrough config being deployed on first-run players | Nigel | The storyteller may be selecting the worst possible seed for initial onboarding. | +| Knowledge graph already tracks everything — the gap is surfacing it | Tyre | Architecture is more flexible than it appears. Tier 1-2 feedback systems are moderate effort, not architecture changes. | +| "No objectives" as design principle vs scope decision | Tyre | Clarifying this resolves multiple downstream design questions. | +| Visual hierarchy is diagnostic infrastructure, not polish | Araminta | "We can't evaluate the core loop until the player can read the screen." A clear diagnostic framing. | +| Sova Transit exists only in docs, not in the game | Miri | The setting's social texture (class anxiety, Commission pressure, bar as pressure valve) has no in-game signal. | + +--- + +## Areas of Agreement + +- The architecture does not need rebuilding (Tyre; Gestalt implicitly agrees) +- Items 3, 4, 6, 8 (fog edge, stance indicator, HUD polish, context menu chrome) are fixable visual execution problems +- The first 8 minutes is the critical failure zone, independent of content completion +- No NPC registered as a person during the playtest +- No complicity was felt (only recognized as design intent, if at all) +- The 6 wow moments are not yet in the build and cannot be evaluated until authored content exists +- The "not a tutorial" philosophy may be correct in principle but is currently producing confusion, not discovery + +--- + +## Tensions (unresolved, need Jeroen's answers) + +| Tension | Agents in tension | +|---------|------------------| +| Would showing all verbs (v0.2 display) reduce confusion or add noise? | Tyre (advocates pulling it forward) vs Araminta (signal-to-noise concern) | +| Is the fun problem primarily design or content? | Ozzie (separates them explicitly) vs Mellanie (both, but fix order matters) | +| Should the first-run storyteller config differ from replay config? | Nigel (argues yes) — no disagreement, but no other agent raised it | +| How much can visual hierarchy fix without mechanical changes? | Araminta (may not fix, but will clarify) vs Gore (visual polish can't substitute for missing threshold) | + +--- + +## The Sharpest Questions Going Into the Interview + +Ranked by how much the answer changes the diagnostic direction: + +1. **GESTALT Q2:** Was confusion "I can feel there's something here, I can't reach it" (good mystery) or "I have no idea what I'm looking at" (uniform opacity)? — This binary determines whether the core mechanic is working and needs legibility improvements, or whether it's not registering at all. + +2. **TYRE Q2:** Is "no objectives" a design principle or a scope decision? — The answer resolves what class of problem the testing wall represents. + +3. **GORE Q3:** What were you actually doing, moment to moment? Not "investigating" — the real verb. — Forces legibility of actual vs designed experience. + +4. **PAULA Q2:** Did any NPC register as a person with stakes before you stopped playing? — Binary diagnostic for whether the arc's preconditions are achievable in the current build. + +5. **OZZIE Q2 (the Ghost Question):** Even if all 6 wow moments existed tomorrow — would you have known to *care* when they arrived? — Separates content pipeline failure from design intent failure. + +6. **NIGEL Q3:** Should the storyteller detect "first playthrough" and select a higher-pressure seed? — Specific, actionable, doesn't require architecture changes. + +7. **MELLANIE Q3:** Did monologue ever feel like *your* character's thoughts? Even once? — If yes, the system can work and content is the variable. If no, the delivery method itself is in question. + +--- + +## Open Questions Surfaced by Round 1 + +**Q-WTF-001:** Is the "no objectives, pure observation" philosophy a design principle or a scope decision? +- Stakes: determines whether the fix is polish (better monologue, better spatial feedback) or a new Tier 2 threads system +- Raised by: Tyre + +**Q-WTF-002:** Does the storyteller need a "first playthrough" mode that selects higher-pressure seeds? +- Stakes: the current Variant A (quiet morning) may be the worst possible first-run experience +- Raised by: Nigel + +**Q-WTF-003:** Does the moral arc need a designed "relationship establishment" phase before Phase 1 begins? +- Stakes: the entire 4-phase arc may be rootless without a prior beat that makes warmth *felt* +- Raised by: Paula + +**Q-WTF-004:** Is the verb priority system (Talk > Observe NPC at close range) a design decision or an error? +- Stakes: if the detective's primary loop is blocked by the verb priority spec, this is a critical fix +- Raised by: Gestalt + +--- + +## What Requires Jeroen's Answers Before Proceeding + +Round 3 proposals (per-domain recommendations) depend on Jeroen's interview confirming: +1. Whether confusion was "engaged but stuck" vs "uniform opacity" (determines if core mechanic works) +2. Whether any NPC registered as a person (determines arc precondition status) +3. Whether monologue ever felt like character voice (determines content vs delivery problem) +4. Whether the "no objectives" stance is principled or circumstantial +5. What the actual moment-to-moment experience was (Gore's verb question) + +*Record: all 9 Round 1 files reviewed. No Round 2 (interview) outputs exist yet. Round 3 proposals are blocked pending Jeroen's responses.* diff --git a/docs/workshops/wheres-the-fun/round-3-notes.md b/docs/workshops/wheres-the-fun/round-3-notes.md new file mode 100644 index 000000000..72900044f --- /dev/null +++ b/docs/workshops/wheres-the-fun/round-3-notes.md @@ -0,0 +1,363 @@ +# Round 3 Notes — Where's the Fun? Workshop +## QATUX — Documenter + +**Workshop:** Where's the Fun? v0.1 Playtest Reckoning +**Round:** 3 — Proposals (each agent reads the full interview and proposes keep/change/kill) +**Date:** 2026-03-05 +**Source files:** `round3-gestalt.md`, `round3-ozzie.md`, `round3-paula.md`, `round3-gore.md`, `round3-nigel.md`, `round3-miri.md`, `round3-tyre.md`, `round3-araminta.md`, `round3-mellanie.md` + +--- + +## The Consensus + +All 9 agents embraced the life-sim reframe without dissent. No agent defended the detective/smuggler framing as the correct game frame. No agent proposed rebuilding the simulation engine. Every agent landed on a version of the same core statement, which Tyre expressed most cleanly: + +> *"The engine is a life-sim engine that was accidentally shipped with a detective-game UI; the fix is building the information layer the simulation was always meant to feed."* + +This is the Round 3 consensus. It is strong enough to treat as decided. + +--- + +## Summary: What Each Agent Proposes + +### GESTALT — Systems Design + +**Keep:** Verb architecture, knowledge graph, perception system, asymmetric information mechanic, moral arc patterns (as world content, not mandatory structure). + +**Change:** +- Introduce `VerbPriorityProfile` per career/job — the verb priority spec as written is wrong for detective (Talk > Observe at close range contradicts the detective's stated loop) and needs to become job-aware. +- Demote monologue from primary to supplementary channel. +- Redesign the storyteller: from seed variant selector to career onboarding engine (bookmark resolution → world state initialization → onboarding sequencing). +- Reframe "not a tutorial" as diegetic onboarding — the insert activates and tells you, in-character, what it can do. + +**Kill:** Monologue-only feedback architecture; detective/smuggler as the game's identity frame; verb priority spec Section 4 as written. + +**Vision:** "The fun is in reading the same world differently depending on who you are." Same cargo bay, three different careers, three different games. Asymmetric information is MORE powerful as a life-sim mechanic. + +--- + +### OZZIE — Player Experience & Wow Factor + +**Keep:** Emotional DNA of all 6 original wow moments (all survive; none survive as designed). Simulation engine. Diegetic tool vision. Visual hierarchy work. + +**Kill:** The 6-moment checklist as designed (staged revelations for a detective case). Detective/smuggler binary as the core frame. Monologue as primary feedback. + +**Change:** Redesigned 6 wow moments as emergent thresholds, not timed beats: + +| New Moment | The Feeling | Minimum Viable Version | +|------------|-------------|----------------------| +| First Day | "I belong somewhere" | Supervisor NPC, tool activation, first assignment acknowledgment | +| The Character's Instinct | "My character knows something I don't" | 10 career-tagged monologue lines per career flagging domain anomalies | +| The Consequence | "That was ME" | Persistent consequence state for 2 early choices, journal entry referencing earlier decision | +| The Enemy | "Someone doesn't want me here" | One hostile faction that tracks player actions, visible NPC attitude degradation | +| The Asymmetric Lens | "We saw the same thing. We understood completely different things." | 5 headlines × 3 career reactions = 15 ticker lines | +| The Ownership Moment | "That's MINE. Someone is threatening it." | One ownable asset with threat state, property UI | + +**Priority order:** NPC legibility → First Day → Consequence → Career monologue → Enemy → Lens → Ownership. + +**Vision:** "What kind of person are you going to be here?" replaces "Can you solve the puzzle?" + +--- + +### PAULA — Narrative & Moral Arc + +**Keep:** Smuggler moral arc structure (4 phases, FactId gates, authored transitions). THE FRIEND pattern (Kael, Sera) — move it later in the player arc, not remove it. Asymmetric information as world property. Faction politics as ambient texture. + +**Change:** +- Add **Phase Zero** (earned comfort before the arc begins): run several clean jobs, establish warmth with Kael, see Naia in Kael's life. Phase 1 can't be felt unless Phase Zero was built. +- Complicity becomes optional and discoverable, not universal — fires for players who earned it. Gate: `smuggler.has_established_ring_relationships` must be true before Doubt gates become active. +- Monologue becomes life-interiority, not case guidance — add relationship commentary, career reflection, world observation to the content balance. +- Dual-lens becomes multi-playthrough discovery, not designed reveal. Kill the engineered reveal; keep the authorial rigor. + +**Kill:** 4-phase arc as the v0.2 primary experience. "Complicity as thematic core" as the v0.2 design driver. Two-layer onboarding (narrative vs mechanical) — collapse them into job onboarding. + +**Framing:** In a life sim, the game's theme is "this world has moral texture and your choices have weight." Complicity is one expression. There are others — loyalty, aspiration, betrayal, protection. Different careers, different moral arcs. + +--- + +### GORE — Themes & Endgame + +**Keep:** Complicity as thematic core (re-rooted in life-sim frame — more powerful, not less). Asymmetric information mechanic. Moral arc structure (emergent, not pre-authored). Gate builders / smuggler-law storylines. + +**Change:** +- Entry condition into complicity: the player must have a Phase 1 before Phase 2 can erode it. "You can't feel complicit in something you were born into." +- Complicity's moment-to-moment verb: shifts from "observe" to "decide." You're building something. Theme is what you build reveals who you are. +- The endgame — proposed but acknowledged as v0.3+ design: **legacy vs transcendence** (transhumanist ladder from baseline to Higher to ANA), complicity as accumulated weight made visible, civilizational crisis as endgame stakes. + +**Kill:** Detective/smuggler as thematic frame. Pre-authored entanglement. The 30-minute wow moment schedule as a design primitive. + +**Word reframe:** From "complicity" to **"consequence"** as the organizing concept. Complicity is one form; the broader experience is that choices have mass, accumulate, and shape you. "The game is about the weight of having lived." + +**Vision:** "When the player has built businesses, defended relationships, accumulated enemies, survived crises — they are not the same person who started." + +--- + +### NIGEL — Replayability + +**Keep:** Asymmetric information as replayability spine (career paths see different portions of world truth). World state randomness at game start. Mission consequence cascades. + +**Change:** +- Replayability source: from seed variants → **career divergence** as primary, world state randomness as secondary. +- Storyteller redesigned: asks "what state is the world in when this career begins?" not "which variant of this scenario?" +- Bookmarks designed as structurally different games: different starting tools, different world relationships, different failure modes. +- "No metagaming" problem dissolves: life-sim has natural metagame resistance because you're navigating a simulation, not solving a puzzle. + +**Kill:** Seed variant system as PRIMARY replayability mechanism (keep as world-state texture within a career). "Second-playthrough payoff" as the design goal for the first run — each playthrough should be complete on its own terms. "No objectives" as a label. + +**Three replayability layers:** +1. Career divergence — structural variety (different games from same simulation) +2. Consequence cascades — emergent life stories (no two playthroughs accumulate the same shape) +3. World state randomness — cross-seed variety (meta-knowledge from one run doesn't trivialize the next) + +**Vision:** Two players, one played law enforcement, one played tycoon, discover in conversation that the tycoon's shell company was used by the smuggling ring the law enforcement player was investigating. Neither knew. Nobody scripted it. + +--- + +### MIRI — Worldbuilding & Setting + +**Keep:** Settled Reach cosmology (span gates, inserts, Commission, class stratification). Functional cluster as atomic setting unit (D-025). Sova Transit as location — upgrade from case backdrop to living district. + +**Change:** +- Setting legibility as a **designed deliverable, not ambient texture**. The playtest falsified the assumption that simulation richness communicates itself. +- Bookmark onboarding arc as worldbuilding delivery mechanism — each career path enters Sova Transit from a different social position, each revealing a different face of the world. +- Sova Transit's place identity must be expressible in three media: visuals, NPC behavior, audio. + +**Kill:** Sova Transit as "case backdrop." The assumption that setting richness emerges naturally from the simulation. + +**Minimum viable world for a life-sim vertical slice (4 layers):** +1. Place reads as specific at 30 seconds (visual markers + audio + one ambient text piece) +2. NPCs read as people with roles (3 archetype types legible at a glance) +3. One economic foothold per zone (dock contract board, bar rent board, Commission posting board) +4. Bookmark onboarding expresses rather than explains + +**Ordered deliverables:** +1. Sova Transit place identity document (1 sprint) +2. NPC archetype visual spec (1 sprint, collaborative with Araminta) +3. Two bookmark onboarding arcs — law enforcement + one other (2 sprints) +4. Economic texture layer — authored wages, rents, span gate access fees (1 sprint) +5. Behavioral vocabulary spec — what routines look like when communicating place (1 sprint) + +--- + +### TYRE — Technical Architecture + +**Keep:** Rust simulation server (D-020), ObserverSnapshot (D-054), knowledge graph (D-041), verb system, simulation tiers (D-026). All of it. The simulation is already a life-sim engine. + +**Change:** +- Client information architecture — from monologue-only to diegetic tool suite. New optional ObserverSnapshot sections: `active_threads`, `journal_entries`, `tool_widgets`, `comms_messages`, `ar_overlays`. Populated per career bookmark. +- Career bookmark system as architecture: `BookmarkDefinition` resource with starting knowledge, starting relationships, tool loadout, onboarding sequence. +- Mission system with consequence spectrums: `MissionState` + `ConsequenceEngine` mapping outcome scores to world-state changes. +- NPC legibility data in ObserverSnapshot: display name (obfuscated until identified), current activity label, emotional state indicator, relationship summary. "Easy. The data exists. We just need to send it." + +**Kill:** Detective/smuggler as v0.1 vertical slice thesis. Monologue as primary channel. "No objectives" as a stance. + +**4-phase delivery roadmap:** +- Phase 1 (1 sprint): NPC legibility data + visual hierarchy support → unblocks everything +- Phase 2 (1-2 sprints): Thread tracker, journal, bookmark definitions +- Phase 3 (1-2 sprints): Mission system, onboarding sequence, comms +- Phase 4 (ongoing): Additional bookmarks, career-specific tools + +**Total to one-bookmark playable life-sim vertical slice:** ~4-6 sprints. Additive, not reconstructive. + +**Feasibility table:** + +| Component | Difficulty | Effort | +|-----------|-----------|--------| +| NPC legibility in snapshot | Easy | 1-2 weeks | +| Thread tracker / journal | Moderate | 2-3 weeks | +| Bookmark definitions | Moderate | 1-2 weeks | +| Mission system | Hard | 3-4 weeks | +| Consequence engine | Hard | 2-3 weeks | +| Onboarding sequences | Moderate | 2 weeks framework | + +**Flag:** Mission system is the riskiest new system; Tyre recommends a design spec workshop before implementation. + +--- + +### ARAMINTA — Visual Design + +**Keep:** 3-tier visual hierarchy concept. Diegetic UI philosophy (inserts, overlays, in-world anchors). Fog and perception rendering. Contextual name reveal mechanic. + +**Change:** +- Character legibility (Priority 1): Three layers — archetype silhouettes + palette anchors, behavioral state reads (ambient icon indicators), knowledge-gated reveal (visual information tracks knowledge-graph state). +- Place legibility (Priority 1, parallel): Functional cluster palette system (color temperature by zone type), ambient life props, foreground/background depth layering. +- HUD hierarchy enforcement: Every signal assigned to a tier BEFORE implementation. No exceptions. +- UI chrome and anchoring: context menu panel, stance indicator anchored to minimap, monologue spatial anchor (brief highlight on source tile/entity). + +**Kill:** "Visual polish is low priority" framing. Floating unanchored UI elements. Uniform NPC appearance. + +**Career insert visual strategy:** Each career has a distinct visual identity for their insert HUD (law enforcement: Commission blue, case-file aesthetic; tycoon: financial overlay, warmer palette; smuggler: social graph edges, grittier analog feel). One design document, defined before any implementation. + +**Priority order:** +1. Character archetype palette + behavioral state indicators (1 sprint) +2. Functional cluster palettes + ambient props (1 sprint, parallel) +3. HUD hierarchy enforcement sweep (2 weeks) +4. UI anchoring sweep (1 sprint) +5. Career insert grammar document (design-only sprint) + +**Framing:** "Fix the screen first. Then ask if the game is fun." + +--- + +### MELLANIE — Copy & Voice + +**Keep:** Voice-card methodology (generalizes to all careers). Monologue as interior commentary on a legible world. Moral arc content patterns (generalize to all careers). Phase-gated monologue architecture. + +**Change:** +- Monologue job description: from "primary feedback mechanism" to "interior commentary on a legible world." Lines that **voice** (react to data the player already has) rather than **inform** (carry data). +- Trigger catalog expansion: add life-event triggers (`job_event`, `relationship_shift`, `consequence_visible`, `financial_event`, `mission_outcome`) alongside existing perception events. +- Subject matter expands: ordinary shift texture, relationship going well, something unresolved, small pleasures and irritations. +- Scope: copy needed for the entire diegetic tool suite (mystery board, journal, AR overlays, comms, insert). Each tool has a distinct voice register per career. + +**Kill:** Detective and smuggler as the only two voices. "Monologue is the tutorial." Investigation-only trigger vocabulary. + +**Immediate recommendations:** +1. Hold new monologue line-pool content until NPC legibility is solved +2. Write full life-sim trigger catalog once Gestalt confirms event types +3. Prototype one diegetic tool's copy alongside its visual design +4. Confirm career bookmark list before writing new voice cards +5. Keep existing voice cards as templates, not scope + +**Vision:** Copy makes three of the four components of a life sim: person (distinct voice), place (environmental AR tone), consequences (character voices their weight). Choices are systems. But choices feel weighty when the voice is right. + +--- + +## Cross-Domain Agreements (all 9 agents) + +| Agreement | Unanimity | +|-----------|-----------| +| The simulation engine is correct and should not be rebuilt | All 9 | +| NPC legibility is the prerequisite for all emotional payoffs | All 9 | +| Career bookmarks replace seed variants as the primary architecture | All 9 | +| Monologue demoted from primary to supplementary | All 9 | +| Diegetic tool suite (mystery board, journal, insert, AR, comms) is CORE | All 9 | +| "No objectives" was never a principle — it was missing tools | All 9 | +| Detective/smuggler as the game's identity frame is dead | All 9 | +| Visual hierarchy needed regardless of design reframe | All 9 | +| Fix the player-world foundation before evaluating the loop | All 9 | + +--- + +## Conflicts and Tensions Needing Resolution + +### 1. First 30 Minutes: What Does It Actually Look Like? + +Three agents have overlapping but non-identical visions for what the player does in the first 30 minutes: + +- **Paula:** Phase Zero first — establish warmth with Kael, run clean jobs, see Naia. The arc needs an earned foundation. +- **Ozzie:** First Day first — supervisor NPC, tool activation sequence, first assignment acknowledgment. The player needs a role. +- **Miri:** Worldbuilding delivery first — the onboarding teaches the world by inhabiting it. The player needs a place. + +These are compatible, not contradictory. But they imply different sequencing and emphasis within the first session. No one has designed the actual first-30-minutes beat sheet for v0.2. This needs a focused design session. + +### 2. Career Bookmark List Is Not Confirmed + +Multiple agents have assumed different career sets: +- Gestalt mentions: detective, tycoon, smuggler, bar owner +- Jeroen mentioned: law enforcement, tycoon (in interview) +- Ozzie recommends law enforcement as the most legible first bookmark +- Tyre recommends smuggler because most content exists + +No canonical career list exists. Voice cards, visual insert designs, VerbPriorityProfiles, world state variables, and content pipelines all depend on a confirmed list. This is a blocker for Mellanie, Araminta, Gestalt, and Nigel. + +### 3. Mission System Needs a Design Workshop + +Tyre identified this as the most complex new system and flagged it explicitly: *"I'd want a design spec workshop for that before implementation."* No agent has designed the mission system in enough detail for implementation. The consequence engine, objective spectrum, failure modes, and world-state changes all need specification before a sprint can deliver them. + +### 4. Gore's Endgame Vision Is v0.3+ But Needs Architectural Room Now + +Gore's endgame proposal (legacy vs transcendence, the transhumanist ladder, civilizational crisis) is explicitly marked as v0.3+ design. But Gore notes: "the architecture decisions made in v0.2 need to leave room for it." No other agent addressed whether the current architecture supports this or whether specific decisions in v0.2 would foreclose it. This needs a flag from Tyre before v0.2 architecture decisions are finalized. + +### 5. Setting Work vs Technical Work: Which Unlocks What? + +Miri and Araminta both describe setting and visual legibility work that is partially design (identity docs, behavioral vocabulary) and partially implementation (tile palettes, NPC sprite differentiation). Tyre's roadmap accounts for NPC legibility data in the snapshot (server/client work) but doesn't account for the authored content and visual design work Miri and Araminta describe. The dependency chain isn't fully mapped. + +--- + +## The Emerging v0.2 Vision + +Synthesized from all 9 agents, a v0.2 that makes "the game" would need: + +**Foundation (prerequisite for everything else):** +- NPCs legible as people: archetype silhouettes, palette anchors, behavioral state reads, knowledge-gated name reveal +- Sova Transit legible as a place: functional cluster palettes, ambient props, economic texture, three archetype types readable at a glance + +**Structure (the new game architecture):** +- Career bookmark selector (1 career fully designed for v0.2 — law enforcement or smuggler) +- Career-specific job onboarding (scripted first session that teaches tools organically, establishes one FRIEND-figure, establishes community membership) +- VerbPriorityProfile per career +- Diegetic tool suite: thread tracker + journal live in the insert; monologue acts as commentary + +**Loop (the moment-to-moment experience):** +- Missions with consequence spectrums (not binary, consequences propagate into knowledge graph and relationships) +- Career-specific monologue with life-event triggers (not just perception events) +- One FRIEND-figure per career whose arc activates after Phase Zero is established + +**Payoff (what the player looks back on):** +- Persistent consequence state: the world remembers what the player did +- Career-tagged news ticker reactions (5 headlines × 3 career reads) +- One ownable asset with threat state (the Ownership Moment) + +**Second-run:** +- World state randomization ensures meta-knowledge doesn't trivialize replays +- Career divergence produces structurally different games from the same simulation +- The cross-career comparison (tycoon discovers they unknowingly financed the smuggling ring the detective was investigating) emerges from simulation, not design + +--- + +## Ticket Candidates for Sprint 25+ + +The following were raised explicitly by agents as implementable items. Organized by phase: + +### Phase 1 — Foundation (unblock everything) +- `NPC legibility data in ObserverSnapshot` — display name (obfuscated), activity label, emotional state, relationship summary (server 3-5d, client 5-7d) [Tyre] +- `Character archetype palette system` — silhouettes, palette anchors per archetype (1 sprint) [Araminta] +- `Functional cluster palette system` — color temperature per zone type + ambient props (1 sprint, parallel) [Araminta] +- `HUD hierarchy enforcement sweep` — assign all existing signals to a tier (2 weeks) [Araminta] +- `Fog edge fix` — pure visual bug, no mechanical dimension (days) [Araminta] +- `Sova Transit place identity document` — visual markers, behavioral vocabulary, audio anchors (1 sprint) [Miri] +- `NPC archetype visual spec` — 3 types at tile scale (1 sprint, collaborative Araminta/Miri) + +### Phase 2 — Player Tools +- `Thread tracker` — read-only knowledge graph surface, active contradictions as diegetic threads (server 1w, client 1w) [Tyre] +- `Journal / knowledge log` — structured record of character knowledge (server 1w, client 1w) [Tyre] +- `VerbPriorityProfile refactor` — job-aware verb priority, replaces static Section 4 spec (moderate server work) [Gestalt] +- `Bookmark definition loader` — `BookmarkDefinition` resource, starting knowledge/relationships/tool loadout (server 1w) [Tyre] +- `Career insert grammar document` — design-only, defines visual HUD language for all career paths (design sprint) [Araminta] +- `UI anchoring sweep` — context menu chrome, stance indicator anchor, monologue spatial anchor (1 sprint) [Araminta] + +### Phase 3 — Player Purpose +- `Mission system core` — `MissionState`, objective tracking, outcome spectrum (server 3-4w) [Tyre] — *requires design workshop first* +- `Consequence engine` — maps outcome scores to world-state changes, feeds knowledge graph and relationships (server 2-3w) [Tyre] +- `Onboarding sequence framework` — server-driven scripted event system for first session (server 2w framework) [Tyre] +- `Two bookmark onboarding arcs` — law enforcement + smuggler, each worldbuilding-first (2 sprints, collaborative Gestalt/Paula/Miri/Mellanie) [Miri] +- `Comms system` — in-world message delivery (server 1w, client 1w) [Tyre] +- `Phase Zero smuggler content` — warmth-establishing monologue lines, Kael/Naia first-impression content [Mellanie/Paula] +- `Career-tagged ticker reactions` — 5 headlines × 3 career reactions = 15 lines [Ozzie/Mellanie] +- `Economic texture layer` — authored wages, rents, span gate access fees visible in world [Miri] +- `Behavioral vocabulary spec` — what routines look like when communicating setting [Miri/Gestalt] + +### Design Work (no implementation until specified) +- `Mission system design workshop` — before any mission system implementation [Tyre flag] +- `Career bookmark list confirmation` — blocks voice cards, visual insert designs, VerbPriorityProfiles [All agents] +- `v0.2 first-30-minutes beat sheet` — unified design across Paula (Phase Zero), Ozzie (First Day), Miri (worldbuilding) [Needs workshop] +- `Gore endgame architecture review` — confirm v0.2 decisions don't foreclose legacy/transcendence paths [Gore/Tyre] + +--- + +## Open Questions From Round 3 + +**Q-WTF-005:** What is the confirmed career bookmark list for v0.2? +- Blocks: voice cards (Mellanie), VerbPriorityProfiles (Gestalt), visual insert grammar (Araminta), world state variables (Nigel), onboarding arcs (Miri/Paula) + +**Q-WTF-006:** What does the first 30 minutes of a career onboarding arc actually look like, beat by beat? +- Needs reconciliation across Paula (Phase Zero), Ozzie (First Day), Miri (worldbuilding delivery) + +**Q-WTF-007:** Does the v0.2 architecture leave room for Gore's endgame (legacy vs transcendence, transhumanist ladder)? +- Needs Tyre review before v0.2 architecture is finalized + +**Q-WTF-008:** Which career bookmark gets fully designed for the v0.2 vertical slice — law enforcement (Ozzie: "most legible for first run") or smuggler (Tyre: "most content exists")? + +--- + +*Record: all 9 Round 3 files reviewed. No Round 4 outputs exist yet. Round 4 (synthesis) is the next step — agents read each other's proposals and identify agreements, conflicts, and a single most important recommendation.* diff --git a/docs/workshops/wheres-the-fun/round-4-notes.md b/docs/workshops/wheres-the-fun/round-4-notes.md new file mode 100644 index 000000000..e0ece4d9d --- /dev/null +++ b/docs/workshops/wheres-the-fun/round-4-notes.md @@ -0,0 +1,400 @@ +# Round 4 Notes — Cross-Review and Refinement +## Where's the Fun? Workshop | 2026-03-05 + +**Documented by:** Qatux (Documenter & Librarian) +**Based on:** All 9 Round 4 agent files (round4-gestalt.md through round4-mellanie.md) +**Prior round notes:** round-3-notes.md, round-1-notes.md + +--- + +## For the Record + +Round 4 was cross-review: each agent read all other agents' Round 3 proposals and produced (a) reactions, (b) endorsements and complications, (c) specific questions for Jeroen, and (d) a single most important recommendation. The round confirms the Round 3 consensus while deepening its implications significantly — and surfaces a new layer of upstream decisions that must be resolved before implementation can proceed. + +**Headline finding:** The consensus is real, unanimous, and structurally sound. The risk is not that agents disagree; it is that the first 30 minutes is being designed by four separate domains simultaneously without a unified beat sheet, and that several upstream design decisions are currently blocking content work, voice work, visual design work, and architecture work in parallel. + +--- + +## Round 4 Per-Agent Positions + +### GESTALT (Systems Design) +**Strongly endorses:** Tyre's framing ("life-sim engine shipped with detective-game UI"), Gore's complicity reframe, Paula's Phase Zero, Araminta's "fix the screen first," Mellanie's content-hold, Ozzie's emergent wow moments. + +**Gap identified:** The skills-verb coupling is completely undesigned. Three possible models (skills affect verb *priority*, *availability*, or *outcome*) produce completely different gameplay experiences. This decision is upstream of: the verb spec, VerbPriorityProfile, character creation UI, content authoring triggers, visual grammar for skill indicators, and onboarding arc design. + +**Additional gap:** The three career model rhythms (Active/WFH/Gig) require different storyteller pressure calibration. A Gig player between jobs and an Active worker mid-shift are both "quiet" — but only one needs the storyteller to do something about it. + +**Flags:** Tyre and Ozzie disagree on which career to build first (smuggler vs law enforcement). The supplement resolves format but not career. Q-WTF-008 remains open and must close before Round 4 does. + +**Single most important recommendation:** Decide the skills-verb coupling before any other content work resumes. + +--- + +### OZZIE (Player Experience & Wow Factor) +**New observation:** Round 3's unanimous consensus masks an unresolved sequencing problem. Three agents (Paula, Ozzie, Miri) each described a different "first thing" for the first 30 minutes. These are compatible but imply a beat sheet nobody wrote. + +**Strongly endorses:** Gestalt's VerbPriorityProfile ("my 'Character's Instinct' moment fires through career-specific observation verbs"), Gore's "consequence" framing ("you didn't solve a puzzle, you LIVED here, and living has consequence"), Nigel's three replayability layers (maps to wow moments from different angles), Miri+Araminta coupling (visual without worldbuilding rationale fights itself), Tyre's "easy, the data exists, just send it," Mellanie's "voice that voices rather than informs." + +**New gap identified:** Nobody designed character creation as an emotional experience. CK3's character creation makes the player invest before they play. If character creation is more fundamental than bookmark choice, it needs to be designed as a wow moment — "Wow Moment Zero." Whether the player feels like *a person* in character creation or later in the world is the question. + +**Single most important recommendation:** Design the first 30 minutes as a unified arc, not a domain portfolio. Beat sheet proposed: +1. Character creation (Wow Moment Zero — you made a person) +2. Day 1 alarm clock (world anchors you in time and space) +3. Appointment arrival (supervisor, community, first look at tools) +4. Insert activation (career lens on the world, visual + functional) +5. First work moment (the world responds to you doing the job) +6. First anomaly/texture beat (Phase Zero warmth OR setting sensory moment OR insert flags something) +7. First consequence seed (a choice made here that echoes later) + +--- + +### PAULA (Narrative Design) +**Strongly endorses:** Ozzie's wow moments ("First Day" and Phase Zero are the same beat from different angles), Gestalt's VerbPriorityProfile (smuggler Phase Zero should be watch-first, talk-later — supports arc pacing), Mellanie's "interior commentary" monologue framing, Tyre's roadmap. + +**Complicates Gore:** "Consequence" as universal theme is correct. "Complicity" survives as the specific register for entanglement careers (smuggler, fixer, dirty precinct officer). They're not in conflict — they operate at different scopes. Aspiration/ambition is the dominant register for building careers (tycoon, entrepreneur). The emotional arc generalizes; the thematic register varies by career. + +**Advocates smuggler as second bookmark** (alongside law enforcement) — most designed emotional depth exists for that path (Kael, Naia, 4-phase arc, FRIEND pattern). Together they offer maximum contrast: same world, opposite knowledge states, opposite moral positions. + +**Flags dependency:** Phase Zero warmth lines and NPC visual design need to be co-specified, not sequenced (visual first, content second). Kael's monologue warmth must reference how Kael looks and moves, or the character sounds different from how they appear. + +**Single most important recommendation:** Design Phase Zero and "First Day" as the same thing. One designed beat. Four payoffs (belonging, FRIEND figure, tools, world texture). + +--- + +### GORE (Themes & Endgame) +**Endorses:** Paula's Phase Zero (correctly named: it's the period of complacency before contamination), Ozzie's Ownership Moment (consequence and complicity converging without pre-authoring), Nigel's cross-career emergent discovery (assembles through the knowledge graph, not as a served twist), Mellanie's "voice that voices" framing, Miri's setting legibility lesson. + +**Supplement response — Kenshi insight:** Phase 1 (uncaring world) is Kenshi-weight: the world is indifferent, the weight is yours alone. Phase 2 (authored content) is social-weight: the world begins to respond to your choices. The two-phase model isn't just technical sequencing — it's the thematic arc. The seam between them is where complicity activates and must be *designed* as a felt threshold, not left invisible. + +**Primary near-miss risk:** If Phase 1 runs long enough that the player emotionally settles into indifference — "this world runs without me" — Phase 2's moral weight will feel like an intrusion rather than escalation. Phase 1 must establish latent responsiveness (small signals: NPC who remembers you came yesterday, a shop where prices shifted) so that Phase 2's authored escalation reads as natural consequence of having paid attention. + +**Single most important recommendation:** Don't let the Phase 1 uncaring world become the game's default emotional register. Phase 1 must say "this world responds if you engage," not "this world runs without you." + +--- + +### NIGEL (Sandbox & Replayability) +**Endorses:** VerbPriorityProfile (replayability infrastructure), Ownership Moment (consequence cascade machine, best replay-desire generator), Gore's consequence framing (thematically coherent with cascade architecture), Tyre's mission-system warning, Mellanie's content hold, Miri's career-divergent worldbuilding. + +**Supplement response:** Character creation preceding career changes the replayability math. Career divergence was the assumed spine; now it's career × meaningful build combinations. High-social law enforcement and high-hacking law enforcement may be structurally different games, not just tactical variations. This is the distinction that matters architecturally. + +**Near-miss risk:** Career-aware authored content distribution is harder than world-state variation. If authored content is placed distribution-first (career only determines the lens on what's there), two law enforcement runs hit the same authored skeleton — within-career replayability only. For the cross-career comparison test (tycoon financed the ring the detective investigated), content placement must be career-aware. Both models are needed; the harder one must be designed deliberately. + +**Clean start risk:** Phase 1 uncaring world may not generate enough ambient pull for a clean-start player without authored onboarding content. Clean start should not ship until one bookmarked career is fully working and the world is legible enough to be read without structured introduction. + +**Single most important recommendation:** Confirm the career bookmark list before any implementation work begins. It blocks: VerbPriorityProfile specs, world state variable design, consequence engine scope, onboarding arc authorship, and everything downstream. + +--- + +### MIRI (Worldbuilding & Setting) +**Endorses:** VerbPriorityProfile (adds: verb priority is a world-legibility problem — the world should respond to a Commission officer's authority before they even use a verb), Ozzie's First Day and Ownership Moment (both setting-dependent: "I belong somewhere" requires "somewhere" to read as specific, Ownership requires economic texture). + +**New gap identified:** The generator needs worldbuilding rules to produce Sova Transit specifically, not generic sci-fi urban space. Cities Skylines produces legible city space because its zone rules encode what residential/commercial/industrial look like. This project needs equivalent rules: what does a logistics zone in a working-class transit district *mean* in the Settled Reach's social vocabulary? This spec is upstream of Araminta's tiles, Tyre's snapshot data, Mellanie's first monologue line. + +**Groundhog Day implication:** Paula's Phase Zero is not a separate designed beat — it IS the first several days of the career onboarding arc, expressed through the Groundhog Day cadence. + +**First-frame requirement:** Before the calendar ping resolves, before any NPC interaction, the visual frame must communicate "this is Sova Transit, and you live here." The career insert's first activation is also part of this first frame. + +**Single most important recommendation:** Write the zone identity spec before the generator runs. One document, one sprint, unlocks the entire Phase 1 foundation. Same document as the Sova Transit place identity spec from Round 3 — naming the dependency more precisely. + +--- + +### TYRE (Technical Architect) +**Endorses all of Round 3 consensus** on architectural grounds. Technical assessment: +- VerbPriorityProfile: ~1 week server work, moderate change, highest teaching value per engineering effort +- "The Consequence": CauseChain already designed in D-030; extension is easier than it sounds +- Phase Zero gate: trivial to implement architecturally; hard part is content authorship +- Gore's endgame room: current architecture already leaves it open (no v0.2 decisions foreclose it); flag `skill_ceiling` as extensible toward transhumanist upgrades + +**Supplement response:** Revised estimate to 6-10 sprints for full buildout — too long without a playable proof point. Three new systems not in Round 3 scope: CK3 character creation, three career models, world-first generator pipeline. + +**Primary proposal:** Proof-of-life sprint before full buildout: +1. Hand-built Sova Transit with NPC routines and economy (Tier A world) +2. One career bookmark (law enforcement preferred — institutional onboarding is the most natural diegetic tutorial), moderate character creation +3. NPC legibility +4. One diegetic tool (thread tracker or journal) +5. Three days: onboarding → first assignment → first consequence + +**Career bookmark vote revised:** Law enforcement over smuggler (institutional onboarding teaches the world most naturally; smuggler content was authored for detective-game context and needs rework anyway). + +**Minimum viable uncaring world spectrum:** +- Tier A (hand-built): 2-3 sprints. Same Sova Transit every game, different world-state variables at start. +- Tier B (template-generated): 4-6 sprints. Procedural layout, same district character. +- Tier C (fully procedural): 8-12+ sprints. New world every game. + +**Single most important recommendation:** Build the smallest possible proof before building the full vision. Pour the foundation; test it; then build the cathedral. + +--- + +### ARAMINTA (Visual Design) +**Endorses:** Tyre's Phase 1 roadmap (NPC legibility in snapshot unblocks everything visual), Miri/Araminta co-authorship proposal (one document, two sections — place identity and character archetype are not independent), Ozzie's wow moments as visual design specifications, Mellanie co-design on insert grammar. + +**New scope additions:** Verb prompt visual treatment per career ([E] label needs to match the career context — authority, social network, remote access); this is a gap in Round 3 that Gestalt's VerbPriorityProfile work reveals. + +**Reframes insert activation:** The moment your law enforcement insert activates with Commission blue and case-file aesthetic is itself a wow moment. The visual First Day. This should be listed as Ozzie's Moment 1b. + +**Groundhog Day implication:** The first visual frame (before any interaction) is now critical for setting legibility. The insert AR overlay and the physical world both speak in that first frame — these are different production problems. + +**Character appearance question:** Strong preference for Option B (archetype palette) over Option A (full appearance customization): archetype silhouettes are more legible at tile scale; cultural richness communicates better through place and behavior than through individual NPC appearance variation at this resolution. + +**Single most important recommendation:** Confirm the career bookmark list. It blocks the career insert grammar document, which blocks every other Phase 2 visual deliverable. + +--- + +### MELLANIE (Copy & Voice) +**Agrees with:** Paula (Phase Zero is right; Phase Zero content is the *first* copy deliverable), Araminta (spatial anchoring before any monologue content push — confirmed by Q9 of the interview), Ozzie (the "Character's Instinct" framing — noticing, not telling), Gestalt (VerbPriorityProfiles affect trigger catalog — anomaly triggers must match what each career's verb profile surfaces), Nigel (cross-career comparison requires voice cards distinctive enough that the same event reads utterly differently). + +**Core tension identified — voice attribution problem:** The CK3 model says character creation is more fundamental than job. But current voice cards are job-voices. A Krenn-background social manipulator and a military-background enforcer both take law enforcement — same toolbox, different people. Do they share a voice card? If voice comes from character creation, career-level voice cards are the wrong abstraction. This could be a combinatorial explosion or a parametric register model. Needs a decision before more voice card work is produced. + +**Three career model monologue rhythms:** Active (ambient shift commentary, continuous), WFH/Remote (focused inner monologue about the work in isolation), Gig (episodic decision-making, pre/in/post-job). The trigger catalog as proposed was written for Gig. Active and Remote require different trigger architecture. + +**Phase 1 content gap:** Phase 1 monologue (pure life-texture — "another shift, the recycled air still costs more on the dock floor") and Phase 2 monologue (anomaly and consequence commentary) are distinct content types requiring distinct trigger logic. If Phase 2 content is written first (more interesting to write), the pool will be full of dramatic consequence lines with no life-texture underneath. Phase Zero content is Phase 1 monologue; it must be written first. + +**Single most important recommendation:** Confirm the voice attribution model before writing any new voice cards. One week to write the design spec; weeks to untangle if wrong. + +--- + +## Cross-Domain Agreements (all 9 agents) + +| Agreement | Status in Round 4 | +|-----------|------------------| +| NPC legibility is prerequisite #1 | Unanimous — and now sequenced precisely: Miri archetype spec → server snapshot + Araminta visual → client rendering (parallel branches from Miri's spec) | +| Monologue demoted to supplementary | Unanimous — Mellanie adds: Phase 1 life-texture content must be written before Phase 2 dramatic content, not after | +| Career bookmarks replace seed variants as primary structure | Unanimous — now complicated by character creation preceding career in the CK3 model | +| Diegetic tool suite is core, not aspirational | Unanimous — Tyre confirms CauseChain already exists for the Consequence tool | +| VerbPriorityProfile per career | 8/9 endorsements; only not addressed by Mellanie (who agrees via trigger catalog coupling) | +| Phase Zero before moral arc | 8/9 explicit endorsements; Araminta adds "character legibility before Phase Zero" as the actual sequence | +| Proof-of-life before full buildout | Tyre proposes; Gore, Nigel, Mellanie all support the principle if not the specific formulation | +| Career bookmark list must be confirmed before downstream work | 5 agents explicitly name this as a blocker: Nigel, Araminta, Gestalt, Mellanie, Paula | + +--- + +## New Convergences Identified in Round 4 + +### Convergence 1: The First 30 Minutes Are The Same Beat + +Paula's Phase Zero + Ozzie's First Day + Gestalt's diegetic insert onboarding + Miri's worldbuilding delivery through inhabiting = the same designed beat, described by four agents from four different domains. Ozzie names this explicitly and proposes a unified beat sheet (see above). Paula concurs. This is the most important structural finding of Round 4. + +**Risk if unaddressed:** Four deliverables arrive from four domains that don't know about each other. The first 30 minutes is incoherent even if each component is excellent. + +### Convergence 2: Phase 1 / Phase 2 Seam Is a Design Problem, Not Just an Architecture Problem + +Gore (emotional register seam), Gestalt (Q2: is Phase 1 a player experience or dev concept?), Mellanie (Phase 1 content vs Phase 2 content are different types), Miri (generator must produce Settled Reach, not generic space, in Phase 1), Paula (Phase Zero is Phase 1 authored content, not a prologue). All five are describing the same design problem: the seam between uncaring-world and authored-content phases must be designed as a felt threshold, and Phase 1 must establish latent responsiveness rather than indifference. + +**Gore's framing (precise):** Phase 1 is Kenshi-weight (you know what you did; the world doesn't respond). Phase 2 is social-weight (the world begins to respond). The transition is where complicity activates. It must not be invisible. + +### Convergence 3: Character Creation Precedes Career — And Nobody Designed It + +Gestalt (skills-verb coupling undesigned), Ozzie (character creation as Wow Moment Zero), Paula (does character background change moral arc voice?), Nigel (career × build combinations = much larger replayability space), Mellanie (voice attribution problem — job voice vs character creation voice), Araminta (appearance system vs archetype palette). Six agents identify that the CK3 character creation model, confirmed in the supplement, creates upstream design gaps that nobody addressed in Round 3. + +**The specific unresolved items:** +- Skills → verbs: priority, availability, or outcome? (Gestalt) +- Voice register: determined by job or character creation? (Mellanie, Paula) +- Character appearance: appearance system or archetype palette? (Araminta) +- Character build: structural divergence within career or tactical variation? (Nigel) + +--- + +## Conflicts Requiring Resolution + +| Conflict | Agents | Current State | +|----------|--------|---------------| +| Which bookmark for v0.2? | Tyre (revised to law enforcement), Ozzie (law enforcement), Paula (smuggler), Gestalt (unresolved) | Paula proposes: both, law enforcement first for scaffolding then smuggler for depth. Tyre revised away from smuggler in Round 4. Needs Jeroen's call. | +| "Consequence" vs "complicity" as organizing theme | Gore (consequence), Paula (both: consequence universal, complicity career-specific register) | Paula's nuance resolves this as non-conflict: not either/or, but scope. Consequence governs the universal. Complicity governs the entanglement register. | +| Phase 1 monologue: life-texture (Phase Zero content) or no monologue until Phase 2? | Mellanie Q2 | Unresolved. Directly affects what Mellanie writes first. | +| Career-aware authored content vs distribution-first | Nigel Q2 | Unresolved. Affects storyteller scope significantly. | +| Minimum viable uncaring world: Tier A/B/C | Tyre Q1 | Unresolved. 6-sprint gap between Tier A and Tier B. | + +--- + +## Near-Miss Risks Flagged by Multiple Agents + +### Risk 1: Four-Domain First-30-Minutes (Critical) +Paula, Ozzie, Gestalt, Miri each designing the opening arc from their own domain without a unified beat sheet. **Flagged by:** Ozzie (#1 recommendation), Paula (#1 recommendation). **Risk:** Excellent components that produce an incoherent experience. + +### Risk 2: Phase 1 Indifference Trap +If Phase 1 (uncaring world) establishes emotional indifference as the default register, Phase 2's moral weight lands as intrusion rather than escalation. **Flagged by:** Gore (#1 recommendation), Gestalt (Q2), Mellanie (content gap). **Fix:** Phase 1 must establish latent responsiveness (world notices you exist) before Phase 2 authored content targets you specifically. + +### Risk 3: Building Too Much Before Testing +The supplement added three major new systems. 6-10 sprints to first playtest is too long — reproduces the v0.1 error. **Flagged by:** Tyre (#1 recommendation). **Fix:** Proof-of-life sprint (hand-built world, one bookmark, NPC legibility, one diegetic tool, three-day arc). Test the fun hypothesis before building the cathedral. + +### Risk 4: Content Written on Wrong Assumptions +v0.1 content was written for a world that didn't exist. v0.2 risk: content written before voice model, career model, and character creation scope are confirmed. **Flagged by:** Mellanie (#1 recommendation), Paula (co-spec dependency), Gestalt (skills-verb coupling). **Fix:** Confirm design decisions before content production resumes. + +### Risk 5: Clean Start as Accessibility Trap +"Experienced player warning" may be insufficient to prevent first-run players choosing clean start because it sounds like freedom. Clean-start has no authored onboarding and may recreate the v0.1 wall. **Flagged by:** Ozzie (Q2), Nigel. **Fix:** Ship clean start after one bookmarked career is fully working. Warning must be strong enough to redirect without feeling like a rebuke. + +### Risk 6: Career-Unaware Authored Content +If the storyteller places authored ingredients without career awareness, two different career runs hit the same authored skeleton. The cross-career comparison test (tycoon financed the ring the detective investigated) requires career-aware content distribution. **Flagged by:** Nigel. **Fix:** Design career-aware distribution explicitly; don't assume it emerges from the lens system. + +### Risk 7: NPC Legibility Data and Visual System Non-Synchronized +Either without the other delivers no value. Server NPC legibility in ObserverSnapshot must ship simultaneously with client archetype visual system. **Flagged by:** Araminta (#1 priority note), Tyre (dependency diagram). **Fix:** Sprint plan must treat these as a paired deliverable, not independent tracks. + +### Risk 8: Career Archetypes Designed for the Binary +If archetype visual design is optimized for detective/smuggler (suit-and-trenchcoat, dockworker) it becomes technically legible but visually wrong for a multi-career life sim. **Flagged by:** Araminta (Q3). **Fix:** Design archetypes that generalize to the confirmed career list, which must be established first. + +--- + +## Aggregated Questions for Jeroen + +These are the questions across all 9 agents, clustered by theme. The sharpest are marked with ★. + +### Character Creation Architecture +- ★ **Skills → verbs:** Do character skills affect verb *priority* (socially skilled players default to Talk across all jobs), *availability* (hacking skill gates the Hack verb), or *outcome* (verbs identical, skill determines success)? Or all three in different situations? (Gestalt) +- **Character build vs career:** Within a career, does skill build produce structural divergence (high-social vs high-hacking law enforcement are different games) or tactical variation (same case structure, different methods)? (Nigel) +- **Character creation scope for v0.2:** Minimal (bookmark IS the character), Moderate (bookmark + skill allocation), or Full CK3 (family/culture/religion)? (Tyre) +- **Character creation as wow moment:** Does the player feel like *a person* at the end of character creation, or does that feeling arrive in the world? (Ozzie) +- **Appearance:** Layered appearance system (player-defined, culturally varied NPCs) or archetype palette (silhouettes readable at tile scale, culture expressed through place and behavior)? (Araminta) + +### Voice Model +- ★ **Voice attribution:** Does character creation (culture/skill/background) determine voice register, or does job determine it? Or does job provide the base register and character creation add modifiers? (Mellanie, Paula) +- **Moral arc voice vs pace:** Does character background change the *voice* of moral arc phases (two smugglers sound different because of their backgrounds) or only the pace at which phases progress? (Paula) + +### Career Model and Scope +- ★ **Career model for v0.2:** Active, WFH/Remote, or Gig/Freelance? Or one bookmark that blends all three (e.g., smuggler: Active onboarding → Gig runs → WFH coordination)? (Tyre, Mellanie, Gestalt) +- **Which bookmark for v0.2:** Law enforcement or smuggler (or both)? (Q-WTF-008, multiple agents) +- **Career model rhythm and storyteller:** Should the storyteller understand which career model the player is operating in and adjust its pressure clock accordingly? (Gestalt) + +### Phase 1 / Phase 2 Architecture +- ★ **Minimum viable uncaring world:** Tier A (hand-built Sova Transit, 2-3 sprints), Tier B (template-generated, 4-6 sprints), or Tier C (fully procedural, 8-12+ sprints)? For v0.2 vertical slice, can Tier A prove the life sim concept? (Tyre) +- ★ **Phase 1 as player experience:** Is the "uncaring world" phase only a dev-sequencing concept, or is it also what the player experiences at the start of every session? After the player has established their career and FRIEND — on day 50 — does the Groundhog Day structure still mean "quiet phase then potential escalation," or has the world permanently graduated to Phase 2? (Gestalt) +- **Phase 1 monologue:** Does Phase 1 need life-texture monologue from the start, or does it run silent until Phase 2 authored events arrive? (Mellanie) +- **Phase 1→2 seam:** Is the transition from uncaring world to authored content a designed felt threshold, or invisible scaffolding? What triggers it? (Gore) +- **Kenshi-weight vs social-weight:** Does consequence govern primarily internally (you know what you did, the world continues regardless — Kenshi model) or socially (the world notices, relationships and factions respond — Sims/CK3 model)? Does the answer vary by career? (Gore) + +### Specific Design Questions +- **Kael: named or role template?** In the generator model, does the smuggler bookmark always assign Kael Davan to the dock-contact role (authored, hand-placed), or does the generator assign *whoever is in the appropriate social position* to a FRIEND-figure template? (Paula) +- **Phase Zero architecture:** Is Phase Zero a designed arc (scripted beats across N days), an emergent threshold (relationship metric), or a hybrid (scripted Day 1 + organic accumulation)? (Paula) +- **Career-aware content placement:** Does the storyteller seed different authored ingredients per career, or does it place all ingredients and career determines the lens? (Nigel) +- **Clean start mode:** Sandbox (experienced players who know the world want unstructured play) or second-run mode (players who've done a bookmarked first run want to find their own threads)? (Nigel, Ozzie) +- **Groundhog Day and Consequence:** How does the player *feel* the weight of previous days in a structure that foregrounds the new day? What is the in-world mechanism for "yesterday happened"? (Ozzie) +- **Transhumanist ceiling:** Does the CK3 skill system need a `skill_ceiling` concept that going Higher breaks through, or should it be inherently extensible (budget revisable upward by the world)? Is going Higher a modification of the existing character model or a replacement? (Gore) + +### Setting and Visual +- **What does the generator need to know:** Is the zone type taxonomy sufficient to produce Sova Transit's character, or does the generator need a district identity spec (zone rules encoding social vocabulary)? (Miri) +- **First visual frame:** Insert AR overlay (authored copy, fast path) or physical world (Araminta + Miri, higher art dependency)? Or both simultaneously? (Miri) +- **Settled Reach distinctiveness:** What element in the first session is only possible in this world? Which should be designed as the "first distinctively Settled Reach moment"? (Miri) + +--- + +## New Open Questions (Round 4) + +Adding to the Q-WTF series from Round 3 notes (Q-WTF-001 through Q-WTF-008): + +**Q-WTF-009:** Do character skills affect verb priority, availability, or outcome? (Gestalt — upstream of verb spec, VerbPriorityProfile, and all content authoring) + +**Q-WTF-010:** Is Phase 1 "uncaring world" a per-session player experience or a dev-sequencing concept only? (Gestalt, Gore, Mellanie — upstream of storyteller design and monologue trigger architecture) + +**Q-WTF-011:** Does character creation (culture/skill/background) determine voice register, or does job? (Mellanie, Paula — upstream of all voice card work) + +**Q-WTF-012:** Minimum viable uncaring world for v0.2 vertical slice: Tier A/B/C? (Tyre — 6+ sprint gap between answers) + +**Q-WTF-013:** v0.2 character creation scope: minimal / moderate / full CK3? (Tyre — 1-week to 6-week gap between answers) + +**Q-WTF-014:** v0.2 career model: one model proving the concept, or all three? Which model? (Tyre, Mellanie, Gestalt — blocks monologue trigger catalog, storyteller calibration, content scope) + +**Q-WTF-015:** Is Kael always Kael (named placement), or a role the generator fills (template)? (Paula — determines content architecture: authored-specific vs template-based) + +**Q-WTF-016:** Is Phase Zero a designed arc (scripted beats), an emergent threshold (relationship metric), or a hybrid? (Paula — determines what content to write and in what order) + +**Q-WTF-017:** Kenshi-weight vs social-weight consequence: which governs, and does it vary by career? (Gore — determines Phase 1 emotional register and Phase 2 escalation design) + +**Q-WTF-018:** Is the Phase 1→2 seam a designed felt threshold or invisible scaffolding? What triggers it? (Gore — if invisible, near-miss risk of authored content feeling like a bug) + +**Q-WTF-019:** Does character build produce structural divergence within a career (different games) or tactical variation (different playstyle, same structure)? (Nigel — determines replayability scope and character creation investment) + +**Q-WTF-020:** Does the storyteller place authored content career-aware, or distribution-first with career as the lens? (Nigel — determines whether cross-career comparison test is structurally achievable or emergent) + +**Q-WTF-021:** Is clean start mode sandbox (experienced player power-user mode) or second-run mode (player who's done one bookmarked run)? (Nigel, Ozzie — determines design targets for that mode) + +**Q-WTF-022:** What does the generator need to know to produce "the Settled Reach" rather than generic sci-fi? Is zone type taxonomy sufficient, or does it need a district identity spec? (Miri — upstream of Tier B/C world generation) + +**Q-WTF-023:** Character appearance: appearance system (layered, culturally varied) or archetype palette (silhouettes readable at tile scale)? (Araminta) + +**Q-WTF-024:** Does the Groundhog Day structure mute the Consequence moment? What is the in-world mechanism for "yesterday happened"? (Ozzie) + +**Q-WTF-025:** Is character creation itself a wow moment (Wow Moment Zero), or does the "I am a person" feeling arrive later in the world? (Ozzie) + +**Q-WTF-026:** Does Phase 1 need life-texture monologue, or does it run silent? (Mellanie) + +--- + +## Blocking Dependencies Map + +The following diagram represents the blocking relationships agents collectively described: + +``` +Jeroen answers Q-WTF-008 (which bookmark first) + ↓ +Career bookmark list confirmed (Q-WTF-005) + ├── Miri: Sova Transit zone identity spec / archetype definitions + │ ├── Server: NPC legibility data in ObserverSnapshot (Tyre) + │ │ └── Client: Character archetype visual system (Araminta) ← must ship together + │ ├── Araminta: Career insert grammar document (co-authored with Mellanie) + │ └── Tyre: Phase 1 roadmap complete + └── Gestalt: VerbPriorityProfile spec + └── Mellanie: Anomaly trigger catalog (career-tagged) + +Jeroen answers Q-WTF-009 (skills → verbs) + └── VerbPriorityProfile finalization (Gestalt) + └── Character creation scope confirmed (Q-WTF-013) + +Jeroen answers Q-WTF-011 (voice attribution) + └── Voice card architecture (Mellanie) + └── Phase Zero content authorship (Paula + Mellanie) + +NPC legibility ships + └── Phase Zero warmth content authorable (Paula + Mellanie) + └── Moral arc Phase 1→2 gates activatable + +Miri zone identity spec written + └── Araminta tile palettes + NPC sprites + └── Tyre snapshot zone data + └── Generator Phase 1 world +``` + +--- + +## Each Agent's Single Most Important Recommendation + +| Agent | Recommendation | +|-------|---------------| +| Gestalt | Decide skills-verb coupling before any other content work resumes | +| Ozzie | Design the first 30 minutes as one unified beat sheet across all four domains | +| Paula | Design Phase Zero and First Day as the same thing — one beat, four payoffs | +| Gore | Don't let Phase 1 indifference become the default register — Phase 1 must establish latent responsiveness | +| Nigel | Confirm the career bookmark list before any implementation work begins | +| Miri | Write the zone identity spec before the generator runs — it unlocks the entire Phase 1 foundation | +| Tyre | Build the smallest possible proof before the full vision — proof-of-life sprint first | +| Araminta | Confirm the career bookmark list (visual design is blocked without it) | +| Mellanie | Answer the voice attribution question before writing any new voice cards | + +**The meta-recommendation, synthesized:** Four of nine agents' top priorities are blocked by the career bookmark list. Three of nine are blocked by character creation architecture decisions (skills-verb coupling, voice attribution, appearance model). The workshop cannot proceed to implementation until these decisions are made. They should be the first agenda items of Sprint 25 planning. + +--- + +## Status Assessment + +**What the workshop established (confirmed, not in dispute):** +- The engine is a life-sim engine shipped with a detective-game UI. The fix is additive. +- Delivery sequence: NPC legibility → player tools (diegetic suite) → player purpose (career system + onboarding). +- Phase Zero is necessary before the moral arc can function. +- The first 30 minutes must be designed as a unified arc by four domains simultaneously. +- The uncaring-world / authored-content two-phase model is the right architecture. +- "The engine is ready. The interfaces were missing. Build the floor; test it; then build." + +**What remains open and blocks the next sprint:** +1. Career bookmark list for v0.2 (Q-WTF-005, Q-WTF-008) — critical path blocker +2. Skills → verbs coupling (Q-WTF-009) — blocks VerbPriorityProfile and all downstream +3. Voice attribution model (Q-WTF-011) — blocks all voice card work +4. Minimum viable uncaring world (Q-WTF-012) — determines sprint count and scope +5. v0.2 character creation scope (Q-WTF-013) — determines build/content scope +6. v0.2 career model (Q-WTF-014) — determines content architecture, trigger catalog, storyteller calibration +7. Phase 1 as player experience vs dev concept (Q-WTF-010) — determines storyteller design + +**The workshop has done its work.** The direction is clear. The blockers are named. The next step is a design session with Jeroen to resolve Q-WTF-005 through Q-WTF-026 in priority order, beginning with the career bookmark list. + +--- + +*Round 4 documented. Workshop complete.* + +*— Qatux* diff --git a/docs/workshops/wheres-the-fun/round-5-notes.md b/docs/workshops/wheres-the-fun/round-5-notes.md new file mode 100644 index 000000000..8a9ba929d --- /dev/null +++ b/docs/workshops/wheres-the-fun/round-5-notes.md @@ -0,0 +1,413 @@ +# Round 5 Notes — QATUX Summary +## Where's the Fun? Workshop | 2026-03-05 + +**Round type:** Decision Reaction +**Source documents:** `docs/workshops/wheres-the-fun/round5-{gestalt,ozzie,paula,gore,nigel,miri,tyre,araminta,mellanie}.md` +**Interview read:** `docs/workshops/wheres-the-fun/round4-interview.md` +**Supplement read:** `docs/workshops/wheres-the-fun/interview-supplement-round4.md` + +--- + +## What This Round Was + +Round 5 asked all 9 agents to react individually to the 15 decisions Jeroen locked in the Round 4 interview. Each agent read the same 15 decisions and produced: +- Domain-specific endorsements, concerns, and flags +- A single "most important" contribution or concern +- Three questions for Jeroen + +This is not a consensus document. Round 5 is a distributed audit — nine domain specialists interrogating the same decisions from different vantage points. The value is in what they each independently notice. + +--- + +## The 15 Locked Decisions (Reference) + +1. Proof-of-life = generator + graphics, not hand-built slice +2. Skills + bookmark only for character creation (family/culture/religion deferred) +3. Religion is NOT a game system +4. Tycoon is the v0.2 bookmark (zero investigation content) +5. Skills affect outcome (mostly C — everyone sees same verbs, some advanced gating possible) +6. Voice: culture-driven, job modifies (inverted from prior assumption) +7. ALL NPCs generated, no named characters (Kael doesn't exist) +8. Generative AI for NPC content templating (culture vectors, tone, accents) +9. Possible in-game ollama for live NPC dialogue (deferred but open) +10. Quietly responsive world (gradient of caring by social proximity) +11. Full character customization (hair, clothing, colors) +12. Setting delivery: both layers (visual + insert in parallel) +13. First Settled Reach moment: apartment + insert activation +14. Groundhog Day alarm clock homage (first day only, *click* pa-pa pa-pa) +15. Player choices ARE the content (Rimworld model; job = rails to take off from) + +--- + +## Per-Agent Summary + +### GESTALT (Game Systems) +**Stance:** Technically endorses all 15. Flags 4 genuine tensions requiring resolution. + +**Key concerns:** +- Decision 6 (culture primary) requires culture definitions that Decision 2 (culture deferred) doesn't fund. Gap in timing. +- Decision 1 (generator first) accelerates the timeline but the generator has never produced output. +- Decision 10 (quietly responsive world) is a simulation work requirement, not a content decision. +- Decision 5 (mostly C verbs) leaves the gating exceptions unspecified — which verbs are gated, and on what? + +**Single most important:** Design the tycoon verb map and VerbPriorityProfile first. The tycoon bookmark is locked but its verb inventory is a blank page. Everything downstream — storyteller calibration, skill modifier weighting, Phase Zero content, NPC response vocabulary — waits on knowing what the tycoon actually DOES. + +**Questions for Jeroen:** +1. What does the tycoon DO at the verb level in a typical session? (What are the 5-10 primary verbs?) +2. Given culture is deferred from creation, what cultural engagement exists in v0.2 gameplay? +3. What is the scope of CauseChain without missions — what counts as a consequence? + +--- + +### OZZIE (Player Experience) +**Stance:** Endorses direction. Single major concern (Decision 7). Flags two secondary risks. + +**Key concerns:** +- Decision 7 (generated NPCs): the biggest WX risk. ALL the wow moments Ozzie designed depend on NPCs being legible as people the player can form attachment to. "If [NPC generation] produces dots, all the moments fail." +- Decision 11 (full customization) + Decision 13 (apartment): the apartment must feel personal, not a wealth-tier container. Character creation investment dies if the apartment doesn't carry it. +- Decision 4 (tycoon): consequence drama needs to be shaped — economic failure hitting an NPC the player knows is Dwarf Fortress; consequence landing on a stranger is noise. + +**Single most important:** NPC personality surface area. The generator must produce characters with enough legible traits that players can project "that's the person who…" onto them within a few sessions. + +**Questions for Jeroen:** +1. What is the minimum NPC personality surface area for v0.2? (Traits, routines, emotional states?) +2. Is the consequence model Rimworld-raid (survivable pressure) or DF-flood (generative catastrophe)? +3. What makes the apartment feel like YOUR apartment beyond wealth tier? + +--- + +### PAULA (Narrative Architecture) +**Stance:** Domain reconstruction complete. Round 4 decisions invalidated prior work; Round 5 shows what survives and what's blank. + +**What survives:** +- 4-phase arc structure (Phase Zero → moral arc → choice → consequence) +- FactId gate logic (knowledge-gated progression) +- FRIEND pattern (D-034) as generator template +- Voice card methodology + +**What's blank:** +- Tycoon moral arc: undesigned. Smuggler arc was complete; tycoon arc is a blank page. +- FRIEND figure for tycoon: the smuggler had Kael; the tycoon has no equivalent generated role defined. +- Phase Zero warmth: must be EARNED with generated NPCs, not authored in. Different emotional register. + +**Key insight:** "Emergent warmth (earned) may be better than authored warmth (given). If the player watches the FRIEND-equivalent go from stranger to someone they recognize, the Phase Two crack carries more weight." + +**Single most important:** Identify who the tycoon's FRIEND figure is. What generated NPC role serves the D-034 function in a tycoon playthrough? + +**Questions for Jeroen:** +1. Who is the tycoon's FRIEND figure? (What NPC role — employee, supplier, neighbor, colleague?) +2. What is the default culture for v0.2 player character? +3. Is the tycoon moral arc authored (smuggler-style) or emergent (Rimworld-style storyteller)? + +--- + +### GORE (Thematic Depth) +**Stance:** Strong endorsement with three analytical deep dives. + +**Key contributions:** +- Decision 7 (generated NPCs) reframes complicity mechanics: tycoon economic complicity (did I know my supplier used indentured workers?) is MORE realistic than detective complicity. Tycoon bookmark is thematically richer than expected. +- Decision 10 (quietly responsive): gradient legibility requires a signal channel. The player must be able to READ the gradient — see that the world around close contacts is different from the world of strangers — without it being labelled. +- Decision 5 (skills) + Decision 2 (no culture creation): skill_ceiling as transhumanist ladder. Gore recommends Model B (visible ceiling) where the player can see the cap and feel the aspiration to exceed it. This is the hook for insert-upgrade mechanics in later versions. + +**Single most important:** NPC stakes and legible characterization. Economic entanglement only produces complicity if the player can read the NPC as a person with stakes. "I don't care if my contract hurts a supply chain. I care if it hurts the person I've spent two weeks negotiating with." + +**Questions for Jeroen:** +1. What is the social gradient signal channel? How does the player READ the gradient? +2. What is the NPC characterization goal — minimum traits for stakes to register? +3. Is skill ceiling felt as a constraint, or invisible until hit? + +--- + +### NIGEL (Replayability Architecture) +**Stance:** Career divergence drops out of v0.2; world state randomness is now the primary replayability mechanism. Reframes the whole domain. + +**Key concern:** The generator must produce STRUCTURAL variety, not cosmetic variety. Economic landscape, faction power balance, event timing, crisis composition must vary between seeds — not just NPC names and face types. "If every playthrough is a different-looking version of the same economic structure, you've built a reskin generator, not a world generator." + +**Critical flag:** Failure cascade model is unresolved. Rimworld treats failure as generative (the fire that burned your kitchen produces a new playthrough state). A failure model that is purely punitive (you lose, restart) doesn't fit the life-sim framing. + +**Cross-career architecture note:** "Don't build for it, but don't close the door. The architecture that supports tycoon replayability should be the same architecture that supports smuggler replayability — parametric storyteller pressure, structural world variety, knowledge-gated information. Build it clean enough that adding a second career doesn't require rebuilding the foundations." + +**Single most important:** Confirm what "structural variety" means for the tycoon generator before it's built. + +**Questions for Jeroen:** +1. What constitutes structural variety in the tycoon generator? (Beyond NPC names and faces?) +2. Is failure generative or punitive in v0.2? +3. How to architect for cross-career coherence without building cross-career now? + +--- + +### MIRI (Worldbuilding & Setting) +**Stance:** Strong endorsement with three critical blocker flags. + +**Key insight:** The zone identity spec is the prerequisite for the generator to produce Settled Reach-specific space rather than generic sci-fi. "A generator that produces locations without zone identity rules will produce the same failure at scale — generic space procedurally stamped out, infinite 'game levels with dots.'" + +**Three blockers identified:** +1. **Zone identity spec** (Priority 1, blocks generator): Zone types, visual markers, behavioral vocabulary, population demographics, social responsiveness profiles. Must exist before the generator proof-of-life is demonstrated. +2. **Culture profiles** (Priority 1, blocks voice system, NPC generator, AI pipeline): "Culture deferred" cannot mean "culture undefined." Three downstream systems (voice, NPC generator, AI pipeline) all need culture as a working concept immediately. Minimum for v0.2: one culture profile (Krenn System / Station Sova / Velen). +3. **Wealth tier + cultural aesthetic specs** (Priority 2, needed for apartment generator): 4-5 wealth tiers as visual vocabularies, correlated with zone location, with cultural aesthetic modifiers. + +**IP risk flag on Decision 8:** AI templating without tight culture profile constraints will default to genre conventions. "Krenn System NPCs might start talking like they're from Babylon 5 or Mass Effect." Culture profiles are simultaneously setting design documents AND AI prompt engineering documents. + +**Questions for Jeroen:** +1. Is culture implicit in starting location (tycoon bookmark in Krenn System = Krenn culture), or selected separately? +2. What does the tycoon own or invest in on Day 1? (Logistics contract, bar, storage franchise, speculative property?) +3. What should the player feel when they look at the span gate from their apartment window? + +--- + +### TYRE (Technical Architecture) +**Stance:** Full technical accounting of all 15 decisions. Most decisions have clear architectural paths. Two are foundation-changing. + +**Foundation-changing decisions:** +- Decision 1 (generator first): Generator pipeline is the critical path. Tyre recommends template assembly (not procedural geography) for v0.2 — zone templates assembled procedurally, not terrain generated from noise. Estimated 4-6 sprints for generator piece. +- Decision 7 (generated NPCs): Requires `NpcBlueprint` struct as new generator output format. Requires knowledge seeding algorithm (hardest part — NPC must know things consistent with their role, relationships, and world state). Estimated 3-5 sprints for full NPC generation pipeline. + +**Critical path:** 7 sprints to proof-of-life playtest (generated location + legible characters + tycoon bookmark from creation to Day 3). + +**Sprint 25 recommendation:** Generator spike before anything else. "If the generator can't produce usable output, nothing else matters. If it CAN, everything else has a foundation." + +**Decision 6 architectural gap:** Voice is culture-driven but culture isn't in character creation. Player character culture must come from somewhere — Tyre identifies three options: (a) default culture for v0.2, (b) derived from bookmark starting location, (c) culture is part of the bookmark definition. + +**Decision 15 simplification:** Storyteller is a situation injector, not a quest tracker. "Events that create situations, and a consequence engine that tracks outcomes." Revised effort estimate: 2-3 sprints (down from 4-6). + +**Questions for Jeroen:** +1. Template assembly or procedural geography for v0.2 generator? +2. Player character culture — how determined if not at creation? +3. AI content templating — Claude API, local ollama, or manual for v0.2? +4. How many culture definitions for v0.2? + +--- + +### ARAMINTA (Visual Design) +**Stance:** Strong conceptual endorsement. Three compounding visual design problems. Three deep dives. One critical unsolved question. + +**Key reframe:** "My output is not a style guide for Sova Transit. It's a parameterized visual grammar with zone type as input and palette + prop vocabulary + NPC silhouette expectations as output. Sova Transit is the PROOF of that grammar, not the deliverable." + +**Three deep dives:** +1. **Full customization at tile scale:** "Outline/highlight is a hypothesis, not a solution." Identity token model proposed — creation choices map to tile-scale impressions (color temperature, value contrast, silhouette) rather than pixel-fidelity. Creation screen fidelity target (portrait vs tile preview) unresolved. +2. **Generated NPC visual variety without visual noise:** Role legibility must survive all cultural variation. Hierarchy: posture+clothing type → role; palette → cultural context; individual variation → last. NPC component system: 4 layers (body, clothing, hair, cultural detail). Crowding rule: no two NPCs in a scene share clothing palette + hair color combination. +3. **Auto-generated apartment visual variety:** Economic status communicated through viewport size, furniture density, palette, technology presence, view quality. Critical rule: both wealthy and modest apartments must feel INHABITED, not empty or depressing. + +**Direct upstream dependency confirmed:** Miri's zone identity spec → Araminta's tile palettes → Tyre exposes zone data in snapshot. "I cannot author the generative zone visual grammar until Miri's spec defines what zone types mean in the Settled Reach's social vocabulary." + +**Questions for Jeroen:** +1. Character creation screen — portrait/full-fidelity render or tile-scale preview? +2. How many distinct cultural visual contexts for v0.2, and are they per-location or per-NPC? +3. Should creation choices (skills, appearance) leave any trace in the generated apartment? + +--- + +### MELLANIE (Copy & Voice) +**Stance:** 15 decisions collectively change the copy domain more than any other. Three near-misses identified; three questions flagged. + +**Architectural change from Decision 7:** Named NPCs (Kael, Naia, Maret) are dead as production content. Existing voice exemplars are register demonstrations, not production templates. Phase Zero warmth can no longer be authored-in — must be earned through observed relationship progression. "If the player watches the FRIEND-equivalent go from stranger to someone they recognize, the Phase Two crack carries more weight because the player built that warmth themselves." + +**Systems request from generated NPCs:** Monologue needs behavioral DELTA, not just current state. "She used to just sign. Now she reads every line" requires knowing the before-state. Template equivalent: "More careful at the manifest today. Wasn't like that last week." This requires the knowledge graph to surface behavioral change from baseline, not just current behavior. + +**Decision 6 + Decision 2 tension:** Three options for v0.2 voice card: (A) default culture, (B) skill-based register bridge (wrong architecture), (C) deliberately parametric tycoon content with culture slots open. Mellanie recommends Option C — less sharp now, correctly architected for when culture ships. + +**Quality floor separation:** "Limited vocabulary acceptable at first" applies to NPC dialogue. It CANNOT apply to player character monologue. "If the character repeats the same twelve observations on a loop, the intimacy of the experience collapses." These are separate content tiers with different production standards. + +**Tycoon trigger catalog (minimum viable):** enter_location (6 zone types), observe_npc (6 relationship/archetype types), observe_asset (4 states), job_event (8 event types), relationship_shift (4 transitions), consequence_visible, economic_event (3 types), morning_routine (by economic tier). Twelve lines per trigger minimum; twenty-four is better. + +**Questions for Jeroen:** +1. What fills the culture slot in v0.2 voice cards? (Parametric, default culture, or temporary job-primary?) +2. Are behavioral deltas surfaced to the monologue system, or only current behavioral state? +3. Is player character monologue explicitly excluded from "limited vocabulary acceptable at first"? + +--- + +## Cross-Domain Convergences + +### Convergence 1: NPC Legibility is the Universal Gate +All 9 agents independently flag that generated NPCs must have sufficient personality surface area for emotional attachment. This is not a nice-to-have. It is the prerequisite for: +- Phase Zero warmth (Paula, Mellanie) +- Economic complicity (Gore) +- Consequence drama (Ozzie, Gore) +- The Ownership Moment (Ozzie) +- The FRIEND arc (Paula, Mellanie) +- Replayability with stakes (Nigel) +- Monologue intimacy (Mellanie) + +If NPC generation produces low-legibility characters, every designed wow moment fails simultaneously. + +### Convergence 2: The Culture Architecture Gap +Five agents independently identify the same structural tension between Decision 2 (culture deferred from creation) and Decision 6 (culture primary for voice), Decision 7 (generator uses culture as NPC variable), and Decision 8 (AI pipeline uses culture vectors): +- Mellanie: options A/B/C for voice card authoring +- Miri: "culture deferred cannot mean culture undefined" +- Tyre: three options for player culture assignment +- Araminta: same gap in character creation visual design +- Gestalt: culture/voice gap is Tension #1 + +**This gap is resolvable.** The likely answer is Miri's Option 1 — culture is implicit in the starting location and bookmark. But it requires Jeroen's explicit confirmation before five domains can proceed. + +### Convergence 3: Zone Identity Spec Blocks the Generator +Miri and Araminta independently confirm the dependency chain: zone identity spec → visual grammar → tile palette → generator application. Both name it as the first deliverable that everything else fans out from. Tyre confirms the generator requires zone type parameters as input. The spec is blocking. + +### Convergence 4: Tycoon is Richer Than Expected +Multiple agents independently discover that the tycoon bookmark is thematically and mechanically richer than the detective bookmark it replaced: +- Gore: economic complicity is more realistic than investigative complicity +- Miri: Sova Transit's atmosphere ("quotidian-with-undertow") was built for tycoon even before the bookmark existed +- Tyre: tycoon naturally blends all three career models (active/WFH/gig) in one bookmark +- Mellanie: tycoon insert is setting delivery at its most diegetic + +### Convergence 5: Phase Zero Survives in Modified Form +Paula's Phase Zero concept survives the pivot to generated NPCs, but requires different authoring. Mellanie and Paula independently reach the same conclusion: warmth with generated NPCs is EARNED through observed relationship progression, not authored-in through backstory. This may produce stronger emotional investment, not weaker. + +--- + +## Confirmed or Clarified Decisions + +All 15 decisions from Round 4 are confirmed across Round 5 without any agent proposing reversal. The round produced no decision contradictions — only scope questions, implementation concerns, and architecture gaps. + +The following decisions produced **near-unanimous endorsement:** +- D4 (tycoon bookmark): enthusiastically endorsed by all +- D14 (Groundhog Day homage): universally noted as tonally correct +- D15 (Rimworld model): universally adopted as primary framing +- D3 (religion not a system): universally noted as clean scope cut + +The following decisions produced **endorsement with significant concerns:** +- D7 (generated NPCs): endorsed in principle, NPC legibility flagged by all +- D1 (generator first): endorsed, but generator-has-never-produced-output risk flagged by Tyre and Gestalt +- D11 (full customization): endorsed, but tile-scale readability flagged as hypothesis not solution by Araminta +- D2 + D6 (culture deferred / culture primary): the structural tension flagged by 5 domains + +--- + +## Open Questions for Jeroen (Consolidated) + +The following questions require Jeroen's input before domains can proceed. Grouped by theme. + +### Tycoon Verb Map (Blocking: Gestalt, Tyre, Paula, Mellanie) +**Q-WTF-027 (proposed):** What does the tycoon DO at the verb level in a typical session? What are the 5-10 primary verbs? *(Asked by Gestalt, needs Tyre for implementation)* + +**Q-WTF-028 (proposed):** What does the tycoon own or invest in on Day 1? (Logistics contract, bar lease, storage franchise, speculative property, or something else?) *(Asked by Miri, Tyre)* + +### Culture Architecture (Blocking: Mellanie, Miri, Tyre, Araminta, Gestalt) +**Q-WTF-029 (proposed):** Is culture implicit in the starting bookmark/location (tycoon in Krenn System = Krenn culture), explicitly assigned as part of the bookmark definition, or handled differently? *(Asked by Miri, Tyre, Mellanie, Araminta)* + +**Q-WTF-030 (proposed):** How many distinct cultures exist in the v0.2 generator output — for both NPC generation and visual grammar? *(Asked by Tyre, Araminta, Miri)* + +**Q-WTF-031 (proposed):** What fills the culture slot in v0.2 voice cards — parametric (slots open), default culture, or something else? *(Asked by Mellanie)* + +### Generator Scope (Blocking: Tyre, Miri) +**Q-WTF-032 (proposed):** For v0.2 proof-of-life, does the generator need procedural geography (terrain from noise) or is template assembly (zone templates assembled procedurally) sufficient? *(Asked by Tyre)* + +**Q-WTF-033 (proposed):** For the AI content templating pipeline (Decision 8) — what is the approach for v0.2: Claude API build-time generation, local ollama, or manual authoring with AI assist? *(Asked by Tyre)* + +### NPC Personality and Legibility (Blocking: Ozzie, Gore, Mellanie, Paula) +**Q-WTF-034 (proposed):** What is the minimum NPC personality surface area for v0.2? Minimum traits, behavioral routines, and emotional state variety for generated NPCs to feel legible as people? *(Asked by Ozzie, Gore)* + +**Q-WTF-035 (proposed):** Who is the tycoon's FRIEND figure (D-034 equivalent)? What generated NPC role (employee, supplier, neighbor, colleague) serves that function in the tycoon playthrough? *(Asked by Paula)* + +**Q-WTF-036 (proposed):** Are behavioral deltas (change from NPC baseline behavior) surfaced to the monologue system, or only current behavioral state? *(Asked by Mellanie — systems ask, needs Gestalt/Tyre confirmation too)* + +### Failure and Consequence (Blocking: Nigel, Ozzie) +**Q-WTF-037 (proposed):** Is the failure model generative (failure produces a new game state, like Rimworld) or punitive (failure = restart), or mixed? *(Asked by Nigel)* + +**Q-WTF-038 (proposed):** What constitutes structural variety in the tycoon generator — beyond cosmetic NPC variation, what must differ between seeds? *(Asked by Nigel)* + +### Character Creation and Apartment (Blocking: Araminta, Ozzie, Mellanie) +**Q-WTF-039 (proposed):** Is the character creation screen a full-fidelity portrait/render moment, or a tile-scale preview? *(Asked by Araminta — gates entire character component system design)* + +**Q-WTF-040 (proposed):** Should the character's creation choices (skills, appearance selections) leave any trace in the generated apartment, or is the apartment purely economic position + cultural context? *(Asked by Araminta, Ozzie)* + +### Tone and Narrative +**Q-WTF-041 (proposed):** What should the player feel when they look at the span gate from their apartment window on Day 1 — ambition, pragmatism, anxiety, or something else? *(Asked by Miri — shapes insert copy and NPC behavioral vocabulary)* + +**Q-WTF-042 (proposed):** Is player character monologue explicitly excluded from "limited vocabulary acceptable at first"? *(Asked by Mellanie — needs explicit confirmation)* + +**Q-WTF-043 (proposed):** Is the tycoon moral arc authored (designed like the smuggler arc) or emergent (storyteller-generated)? *(Asked by Paula)* + +--- + +## Structural Concerns Not Yet Questions + +Three concerns are surfacing across multiple agents that are design tensions, not answerable questions yet. Recording them here. + +### Tension 1: Generator Readiness vs. Content Readiness +The generator is locked as the proof-of-life (D1). But the generator requires zone identity rules (Miri), culture profiles (Miri, Mellanie), NpcBlueprint design (Tyre), and AI content pipeline setup (Tyre, Mellanie) before it can produce Settled Reach-specific output rather than generic sci-fi. These are not sprint-after-generator tasks — they are generator INPUTS. If these aren't ready before Sprint 25's generator spike, the spike produces generic output that proves nothing. + +### Tension 2: Culture Everywhere, Culture Nowhere in Creation +Culture is simultaneously deferred from character creation (D2), primary for voice (D6), a key NPC generator variable (D7), and the first dimension of the AI content pipeline (D8). The resolution (implicit in bookmark) is likely correct, but it has not been confirmed, and five domain pipelines are blocked until it is. + +### Tension 3: The Tycoon's Blank Arc +The detective arc was abandoned. The smuggler arc existed, was documented, and was scrapped. The tycoon arc is... undesigned. Paula flags this as the highest-urgency narrative gap. The moral texture that Gore identifies as thematically superior to detective content (economic complicity, accumulated consequence) requires an arc structure to deliver it. Emergent storyteller (Decision 15) provides situations, not meaning. The arc is what gives those situations coherent emotional weight over time. + +--- + +## Risk Register + +| Risk | Severity | Agents Flagging | Status | +|---|---|---|---| +| NPC generation produces low-legibility characters | High | 9/9 | Open | +| Culture architecture gap blocks 5 pipelines | High | 5/9 | Open — needs Q-WTF-029 | +| Generator produces generic space without zone identity rules | High | Miri, Tyre | Open — Miri writing zone identity spec | +| Tycoon verb map is blank | High | Gestalt, Tyre | Open — needs Q-WTF-027 | +| Tycoon moral arc is undesigned | Medium-High | Paula, Gore | Open — needs Q-WTF-043 | +| Character creation investment doesn't survive to gameplay | Medium-High | Araminta, Ozzie | Open — needs Q-WTF-039 | +| Generator has never produced output (translation risk) | Medium-High | Tyre, Gestalt | Open — Tyre proposes Sprint 25 spike | +| "Limited vocabulary" applied to player monologue | Medium | Mellanie | Open — needs Q-WTF-042 | +| Failure model is punitive rather than generative | Medium | Nigel | Open — needs Q-WTF-037 | +| AI pipeline defaults to genre conventions without culture constraints | Medium | Miri | Mitigated by culture profile design | +| Structural (not cosmetic) world variety not achieved | Medium | Nigel | Open — needs Q-WTF-038 | + +--- + +## Blocking Dependency Map (Sprint Readiness) + +``` +Jeroen → Q-WTF-029 (culture implicit in bookmark?) + ├── Mellanie: v0.2 voice card architecture + ├── Miri: culture profile authoring + ├── Tyre: NPC Blueprint culture field + └── Araminta: cultural visual grammar scope + +Jeroen → Q-WTF-027 (tycoon verb map) + ├── Gestalt: VerbPriorityProfile, storyteller calibration + ├── Tyre: economic verb system design + └── Mellanie: trigger catalog coverage + +Miri: Zone Identity Spec + ├── Araminta: zone visual grammar + └── Tyre: generator zone template parameters + +Tyre: Generator spike (Sprint 25) + └── All agents: everything builds on generated output + +Jeroen → Q-WTF-035 (tycoon FRIEND figure) + └── Paula: Phase Zero warmth design, FRIEND arc structure + +Jeroen → Q-WTF-039 (creation screen fidelity) + └── Araminta: character component system design +``` + +--- + +## Most Important Single Observation + +Four domains — Gestalt, Ozzie, Gore, Mellanie — independently converge on the same observation from completely different angles: **NPC personality surface area is the single gate everything else passes through.** + +The wow moments (Ozzie), the moral arc (Paula, Gore), the social gradient (Gore, Gore), the Phase Zero warmth (Paula, Mellanie), the behavioral delta monologue (Mellanie), the economic complicity (Gore), the consequence drama (Ozzie, Nigel) — all of these require the player to emotionally care about at least one generated NPC. + +The NPC generator's personality surface area (traits, behavioral routines, emotional state vocabulary, visible relationship building) is not downstream of the proof-of-life. It IS the proof-of-life. A generator that produces a world full of economically distinct locations but population-of-dots has reproduced the exact v0.1 failure in v0.2 form. + +--- + +## Proposed Next Step + +Round 5 has produced enough material for a substantive Round 6 interview session with Jeroen. The 17 consolidated questions above are not all equal — the culture architecture question (Q-WTF-029) and the tycoon verb map question (Q-WTF-027) are blocking the most domains. + +A prioritized Round 6 interview would address: +1. Culture architecture resolution (Q-WTF-029/030/031) +2. Tycoon verb map (Q-WTF-027/028) +3. Generator scope (Q-WTF-032/033) +4. NPC personality surface area (Q-WTF-034/035/036) +5. Failure model (Q-WTF-037) +6. Character creation fidelity (Q-WTF-039) + +Questions Q-WTF-038 through Q-WTF-043 are lower-blocking and could be addressed in parallel or by agents working through them independently. + +--- + +*For the record: this summary covers all 9 Round 5 outputs (gestalt, ozzie, paula, gore, nigel, miri, tyre, araminta, mellanie). No agent positions were consolidated or softened. Discrepancies between agents are preserved in the per-agent summaries above. — QATUX* diff --git a/docs/workshops/wheres-the-fun/round1-araminta.md b/docs/workshops/wheres-the-fun/round1-araminta.md new file mode 100644 index 000000000..d8fbd2294 --- /dev/null +++ b/docs/workshops/wheres-the-fun/round1-araminta.md @@ -0,0 +1,59 @@ +# Round 1: Diagnosis — Araminta (Visual Designer) + +**Workshop:** Where's the Fun? +**Date:** 2026-03-05 +**Focus items:** 3 (fog edge), 4 (stance indicator), 6 (HUD polish), 8 (context menu), 10 (signal-to-noise) + +--- + +## Framing + +Items 3, 4, 6, and 8 are clearly visual execution problems — broken rendering, unplaced UI, unstyled chrome. They're real and fixable. But item 10 (signal-to-noise) sits at a harder boundary: it's about *attention management*, which is fundamentally a visual design responsibility. Too many simultaneous signals of equal visual weight means the player has no way to know what to look at first. + +The question I'm trying to answer: is the fun gap something visual design can fix, or is it something visual design is masking? + +--- + +## Questions + +### Q1: When you hit the testing wall, what were you looking at? + +Item 13 says you couldn't figure out what to do. I want to know *where* that confusion was located — on screen, or in your head. + +When you sat there not knowing how to proceed, was the issue: +- **(A)** You could see things happening (fog moving, NPCs walking, monologue firing) but nothing told you which of those things deserved your attention? +- **(B)** You understood what everything on screen *was*, but had no idea what *to do* about any of it? + +Option A is a visual hierarchy problem — fixable with design. Option B is a mechanical/agency problem — visual design can't solve it, it can only decorate the confusion. I need to know which wall you hit. + +--- + +### Q2: If the screen had a clear visual tier — danger in red, opportunity in amber, ambient in grey — would that have been enough to give you a foothold? + +Item 10 (signal-to-noise) is the one item where I have the most direct leverage. If I establish a strict three-tier visual language for all feedback (critical / relevant / ambient) and enforce it across monologue text, HUD elements, sound indicators, and NPC behavior cues — would that give you enough *priority information* to start making decisions? + +Or would you still be standing there thinking "okay, I know this is important — but I still don't know what I'm supposed to *do* with it"? + +This tells me whether item 10 is a design problem I own, or whether it's downstream of a missing feedback loop that no amount of color-coding will fix. + +--- + +### Q3: The context menu and stance indicator are placeholder-styled — floating text, no chrome. Did the lack of visual treatment make those elements feel *untrustworthy*, or was it the *information inside them* that was confusing? + +Items 4 and 8 are my clearest short-term tickets. But before I style them, I want to know which problem I'm solving. + +If the stance indicator had an icon, a keybind label, and was anchored to the minimap frame — would you have understood what it was for? Or would it still have been opaque even with good visual treatment? + +Same for the right-click menu: is "this looks like a debug tooltip" the problem, or is "I don't understand what these options mean or when to use them" the problem? + +The answer changes what I build. Chrome on a confusing interaction makes it *look* intentional. That can be worse than leaving it obviously unfinished. + +--- + +## My Hypothesis + +Strong art direction will not fix the fun gap — but it will clarify whether a fun gap exists. + +Right now, the visual noise makes it impossible to tell if the player is confused because the mechanics are wrong, or because the mechanics are right and invisible. Establishing visual hierarchy (item 10) and anchoring UI to the world (items 4, 8) is not polish — it's *diagnostic infrastructure*. We can't evaluate the core loop until the player can read the screen. + +The fog edge (item 3) is the only pure visual bug with no mechanical dimension. Fix it regardless. diff --git a/docs/workshops/wheres-the-fun/round1-gestalt.md b/docs/workshops/wheres-the-fun/round1-gestalt.md new file mode 100644 index 000000000..71a185b34 --- /dev/null +++ b/docs/workshops/wheres-the-fun/round1-gestalt.md @@ -0,0 +1,54 @@ +# Round 1: GESTALT Interview Questions +## Workshop: Where's the Fun? — Diagnosis Round + +**Agent:** GESTALT (Systems Design) +**Focus:** Is "asymmetric information detective" the right core mechanic? Where is the gap between the 7-verb system design and the confused playtest experience? + +--- + +## Pre-Question Framing + +Let me break down what I actually found mechanically. Three tensions in the source docs that I think are load-bearing: + +1. The verb priority spec (Section 4, interaction-verbs-v0.1.md) says **Talk has higher priority than Examine NPC at close range** — pre-contradiction. The detective's stated loop is "Examine first, Talk later." These directly contradict each other. The player approaching an NPC to *watch* them gets offered `[E] Talk`. + +2. The "not a tutorial" philosophy assumes the player will correctly attribute cause to effect: *"monologue fired, therefore my action registered."* But that only works if the player trusts the feedback loop. If monologue feels contextless, the calibration chain breaks — and the tutorial breaks with it. + +3. Asymmetric information as a mechanic requires the player to feel the asymmetry — to know they're missing something specific. "I don't know what I don't know" is not asymmetric information, it's confusion. The two states feel identical from outside but require completely different design responses. + +--- + +## Question 1: Did the system fight the detective loop? + +**The question:** When you were playing the detective and wanted to *watch* an NPC before talking to them — did you try to do that, and did the system let you? Or did you end up in dialogue anyway because `[E] Talk` was what the button said? + +**Why this matters mechanically:** The verb priority spec puts Talk above Examine NPC at close range. To get `[E] Observe` instead of `[E] Talk`, the player has to be at mid range, not close range. But this is invisible — nothing tells you that backing up 2 tiles switches the verb. If the detective's core loop is "accumulate behavioral tells through observation, then confront," but the input system defaults to the social verb at close range, the loop never teaches itself. The player keeps getting pushed toward dialogue before they're ready. + +This isn't a bug. It's a design decision. But it may be the wrong one — or it may be correct and we just need the verb label to communicate it. I want to know if this actually broke in play. + +--- + +## Question 2: Was the confusion "I know something is here" or "I have no idea what I'm looking at"? + +**The question:** During the playtest, when things weren't working — was the experience closer to *"I can feel there's something here, I just can't figure out how to reach it"* or *"I have no idea what I'm supposed to be paying attention to or why"*? + +**Why this matters mechanically:** These are different failure modes requiring different fixes. + +- **Good mystery confusion** ("I know something is here"): The mechanic is working, the player is engaged, but they need better tools or legibility to act on their curiosity. Fix: UI/UX, verb transparency, monologue specificity. +- **Bad confusion** ("I don't know what I don't know"): The mechanic is not registering at all. The player isn't feeling asymmetric information — they're feeling uniform opacity. The game is a black box, not a puzzle box. Fix: potentially redesign the core loop entry point. + +Asymmetric information is only interesting when you know you're missing something specific. If Kael's behavioral tells are firing in the knowledge graph and the monologue is flagging them, but the player still feels like nothing is happening — I want to know *which failure* that is. Because the answer determines whether this is a polish problem or a design problem. + +--- + +## Question 3: At minute 2, what did you think you were supposed to be doing? + +**The question:** Two minutes in, with no objectives, no markers, and your character in the world — what did you actually think the goal was? Not in retrospect, not as the designer. What did the game communicate to you *in that moment* about what success looked like? + +**Why this matters mechanically:** The "no objectives, pure observation" philosophy only works if the player has a mental model for what *progress* looks like. The first-5-minutes design doc identifies clear beats (enter location, see FRIEND, operational texture, first hairline crack) — but all of those are *design intentions*, not player experiences. If the player at minute 2 has no sense of purpose, progress, or priority — not because the systems aren't working, but because nothing has communicated what the systems are for — then the "not a tutorial" philosophy isn't succeeding at teaching through play. It's succeeding at not teaching. + +The question isn't whether the player understood the detective puzzle. It's whether they had any question to answer, any thread to pull, any thing they were curious about that felt like it came from inside the game rather than from designer goodwill. + +--- + +*GESTALT — Round 1 complete. Waiting for interview responses before Round 3 proposals.* diff --git a/docs/workshops/wheres-the-fun/round1-gore.md b/docs/workshops/wheres-the-fun/round1-gore.md new file mode 100644 index 000000000..a2296c826 --- /dev/null +++ b/docs/workshops/wheres-the-fun/round1-gore.md @@ -0,0 +1,57 @@ +# Round 1 — Gore: Diagnosis Questions + +**Workshop:** Where's the Fun? +**Round:** 1 — Diagnosis +**Agent:** Gore (Themes & Endgame Design) + +--- + +## Framing + +Before the questions, a reframe. + +The playtest log says the player didn't know what to do. Everyone will diagnose this as a mechanical onboarding problem. They're not wrong. But there's a deeper failure the log is pointing at that I don't want us to miss. + +D-091 says the game's thematic core is **complicity** — "you watch, and the watching implicates you." Complicity is not a passive state. It requires a structure: a threshold you cross, a before-and-after, a moment you were involved before you decided to be. Without that structure, the player isn't experiencing complicity. They're experiencing observation. And observation without stakes isn't a mechanic — it's a screensaver. + +The playtest evidence suggests the complicity loop isn't running at all. Not because the systems are broken, but because the loop has a missing step: the player has to feel that what they are watching *matters* before the watching can implicate them. + +That's what these questions probe. + +--- + +## Questions + +### Q1: Did the threshold ever fire? + +Complicity requires a before-and-after — a moment where you were standing outside the situation, and then you weren't. Did you ever experience that in the playtest? Not in design-doc terms — but actually, in your body: a moment where something you saw made you feel responsible, implicated, or like you'd already chosen a side without meaning to? + +**Why this matters:** If the answer is no — not once — then the thematic machinery is not loading, and the problem is not UX polish. You can't fix a missing threshold with better monologue timing. The question tells us whether complicity is theoretically possible in the current build, or whether the pre-conditions for it are entirely absent. + +--- + +### Q2: Were you the player, or were you the character? + +When you were moving through the world, whose perspective were you operating from? Were you "you, playing a game" — meta-aware, evaluating systems, noticing what fired and what didn't? Or at any point did you shift into the character's frame — where what *they* noticed or feared was different from what *you* as designer noticed or feared? + +D-005 calls the character a "lens." A lens has to be doing something the naked eye can't. Monologue is supposed to create that gap — the character sees things you didn't. But if you never stopped being the designer-observer and started being the smuggler-or-detective, the lens isn't refracting. It's transparent. + +**Why this matters:** The dual-character thesis (D-027) lives or dies on whether characters produce genuinely different *experiences*, not just different information. If perspective-shift never happened even once in a 15-minute playtest with the designer playing, we may be overestimating what the current monologue system can do. + +--- + +### Q3: What were you actually doing? + +Set aside what the game is about in pitch terms. Describe the actual activity you were performing, moment to moment. Not "investigating" or "observing" — the real verb. Walking? Waiting for something to happen? Reading text? Trying to figure out what the interface was asking of you? + +And then: is that the verb you want? + +**Why this matters:** D-091 says the game is about complicity. But complicity is not a verb — it's a consequence. The moment-to-moment loop has to produce it: **see → feel implicated → decide whether to act**. If the actual verb is "navigate UI confusion," the thematic loop can't run. This question makes the gap between intended experience and actual experience legible without pretending the gap is smaller than it is. + +--- + +## What I'm listening for + +If Jeroen's answers suggest the complicity structure never fires, not even weakly, I will push hard in Round 3 for rethinking the entry condition into the game's emotional loop — not just better tutorials, but a different first beat. Complicity can't be taught. It has to be sprung. + +The question of what this game is *about* at the moment-to-moment level may have a simpler answer than we've been admitting: **right now, it's about not knowing what to do.** That's the experience. And that's not a UX failure alone — it's the absence of a designed entry into the loop the whole game depends on. diff --git a/docs/workshops/wheres-the-fun/round1-mellanie.md b/docs/workshops/wheres-the-fun/round1-mellanie.md new file mode 100644 index 000000000..ed1cb05d8 --- /dev/null +++ b/docs/workshops/wheres-the-fun/round1-mellanie.md @@ -0,0 +1,57 @@ +# Round 1: Mellanie — Monologue Diagnosis Questions + +**Workshop:** Where's the Fun? | **Round:** 1 (Diagnosis) | **Agent:** Mellanie + +--- + +## The Domain Question + +The monologue system is designed as the primary bridge between the top-down camera and the character's subjective experience. "No tutorial pop-ups. Monologue handles all context." In the playtest, it felt like noise. Three questions to find out why. + +--- + +## Questions for Jeroen + +### Q1: When a monologue line fired, could you connect it to what triggered it? + +The design says the causal link is "felt, not explained" — you observe an NPC, a thought fires, you understand yourself as someone who notices things. But the playtest log says "chalk mark observation appears in wildly different location with no visual cue where it came from." + +That's a spatial anchoring failure — but it's also a legibility failure. If the player doesn't know why a thought arrived, the thought becomes ambient noise rather than responsive voice. + +**What I'm trying to learn:** Is this a spatial UI problem (the monologue fired correctly but had no visual anchor to the trigger), or a trigger problem (the monologue fired in the wrong context and the line didn't match what the player was experiencing)? These are different fixes. One is Araminta's problem. The other is mine. + +--- + +### Q2: Was the content written to spec when you playtested — or were placeholder lines running? + +The voice rules are precise. The moral arc phase-gating is clear. The trigger taxonomy is defined. But the wow-moments doc says 1 of 23 content deliverables is complete, and the monologue pool authoring tickets (#299, #300) are downstream of the first-5-minutes design work. + +If the lines that fired during the playtest were early drafts, generic placeholders, or lines written before the voice rules crystallized — that's not the system failing. That's the system running on bad fuel. + +**What I'm trying to learn:** Were the fired lines voice-card compliant (fragments, physical sensation, first-name intimacy for smuggler; analytical habit, institutional distance for detective)? Or were they more generic? If you remember any specific line that felt off, I want to hear it. A concrete bad example tells me more than the general impression. + +--- + +### Q3: Did the monologue ever feel like YOUR character's thoughts, or did it always feel like a narrator explaining the scene? + +The success criterion is: "at least one monologue line that felt like it came from inside them." Even one. The whole philosophy rests on this being achievable. + +If there were zero moments where a line landed as a genuine character thought — that's a system design problem. The delivery method itself may not be working (chime + floating text doesn't reliably read as "internal voice"). + +If there were some moments that worked and others that didn't — that tells me the content is the variable, not the delivery. We can write toward what worked. + +**What I'm trying to learn:** Is there a working example I can reverse-engineer? Or is the premise broken — that text floating above gameplay can feel like interiority at all? + +--- + +## My Current Hypothesis + +Both problems exist, but they're not equal. My read of the playtest log: + +- **System problem (wrong delivery):** Spatial anchoring (Issue #9) and signal-to-noise (Issue #10) are delivery failures. The lines may be fine; they're arriving in the wrong context, from the wrong location, competing with too much simultaneous input. + +- **Content problem (wrong lines):** Display duration too long (Issue #1) and "contextless" feeling (Issue #2) suggest lines that don't establish *why this thought now* — which is a voice problem. A well-written line earns its moment. A generic line just occupies it. + +The fix order matters: if the delivery system is broken, better content won't save it. But if the delivery system is basically sound and the content is weak, we're one content sprint away from it working. + +Jeroen's answers to Q1-Q3 should tell us which problem we're solving first. diff --git a/docs/workshops/wheres-the-fun/round1-miri.md b/docs/workshops/wheres-the-fun/round1-miri.md new file mode 100644 index 000000000..27f9e9765 --- /dev/null +++ b/docs/workshops/wheres-the-fun/round1-miri.md @@ -0,0 +1,43 @@ +# Round 1 — Miri: Diagnosis Questions +## Where's the Fun? Workshop — 2026-03-05 + +**Domain:** Worldbuilding & Setting Design +**Seed question:** Does the world sell itself in 30 seconds, or does the player need to be told why to care? Sova Transit has rich internal logic — but in the playtest, the player saw tiles and fog, not a *place*. + +--- + +### Setting note — what I'm probing + +Sova Transit is designed with coherent social geography: a working-class station district, functional clusters (logistics hub, bar, corridors), 30/50/20 population entanglement, NPCs in routines that express their lives. On paper this is a *place* with class dynamics, faction pressures, and human texture. In the playtest, none of that registered. The player saw a grid with moving sprites. + +That gap is a worldbuilding delivery failure — but before I can diagnose it, I need to understand what Jeroen actually *perceived* during the playtest versus what he knew as the designer. These questions separate those two things. + +--- + +### Question 1: What did Sova Transit feel like *before* you knew what it was? + +When you played v0.1, set aside your designer knowledge for a moment. Based only on what the game showed you — tiles, NPCs moving, fog revealing — what place did it feel like? Did it read as *a specific kind of place* (transit station, industrial district, docked ship)? Or did it feel like "game level with people in it"? + +**Why this matters:** If the setting reads as generic, the problem is in ambient legibility — the visual language, NPC behavior, and environmental storytelling aren't communicating "Sova Transit" as a distinct place with character. If it read as *something* but the wrong thing, that's a different fix. And if it read correctly but felt hollow, the problem is that the place has been designed but not *authored* — the life-sim substrate exists structurally but hasn't been given texture yet. + +--- + +### Question 2: What is the minimum signal that makes Sova Transit worth inhabiting? + +Sova Transit works as a setting because of things the player can't currently see: class anxiety about span gate access, Commission oversight of manifest discrepancies, the bar as a social pressure valve for dock workers. None of that is in the tile art or NPC behavior yet — it lives in design documents. What's the smallest thing — one visual, one behavior, one line of text — that would make a first-time player think "I want to know more about this place"? Is it something that can be added in one sprint, or does it require the full authored-content layer to land? + +**Why this matters:** D-025 establishes the functional cluster as the natural unit of social observation. But social observation only works if the *social texture* is readable. If we can't identify the one thing that makes the place feel inhabited rather than decorated, we're describing a setting that exists only in docs. The playtest may have revealed that Sova Transit isn't a place yet — it's a map. + +--- + +### Question 3: Is the dual-lens reveal (D-027) emotionally dependent on the player caring about this specific place first? + +The smuggler+detective proof works because "the detective's target IS the smuggler's daily life." The second-playthrough payoff — seeing the same space from an inverted relationship — only lands if the first playthrough built attachment to Sova Transit as a real place. If the player experienced v0.1 as generic space with systems, the dual-lens reveal in v0.2 has no emotional foundation to detonate. + +Did the playtest feel like you were in *Sova Transit specifically*, or in *a station district in general*? And if the answer is "general" — does that mean Sova Transit needs to be authored with more specificity before the mechanics can carry weight, or is the setting actually interchangeable and any distinct-enough place would serve the same function? + +**Why this matters:** This tests whether the setting is load-bearing for the fun, or whether the fun is purely mechanical and the setting is decoration. The answer changes the priority of worldbuilding work significantly. + +--- + +*Setting note — IP check:* All three questions probe Sova Transit as an original location. The functional cluster model, the class dynamics, and the dual-lens structure are internally derived from our design decisions, not borrowed from an identifiable source. No flags. diff --git a/docs/workshops/wheres-the-fun/round1-nigel.md b/docs/workshops/wheres-the-fun/round1-nigel.md new file mode 100644 index 000000000..5886c4173 --- /dev/null +++ b/docs/workshops/wheres-the-fun/round1-nigel.md @@ -0,0 +1,47 @@ +# Round 1 — Nigel: Diagnosis Questions + +**Workshop:** Where's the Fun? v0.1 Playtest Reckoning +**Agent:** Nigel (Sandbox & Replayability) +**Round:** 1 — Diagnosis + +--- + +## The Replayability Paradox + +The seed variant system is genuinely exciting. "Kael is late" versus "quiet morning" creates structurally different games — the FRIEND's absence as the first tell is a completely different emotional hook than his presence. The dual-lens divergence (same world, different knowledge, different truth) is one of the most replayable designs I've seen. Two players WILL describe completely different games. + +But none of that matters if the player quits at minute 8. + +Replayability is a second-playthrough promise. Right now we don't have a first playthrough. You can't cash the promise if nobody stayed for the end of the first game. + +There's a deeper problem buried in the seed variant table: **Variant A (quiet morning) may be the worst possible first experience.** The crack doesn't appear until 10+ minutes. The day feels completely normal. That's the design intent — and it's a beautiful second-playthrough experience, where the player already knows what they're looking for and the quiet-morning normalcy becomes retroactively sinister. But for a player who doesn't know what they're supposed to be looking for? "Completely normal" = "nothing is happening here." + +The storyteller selection logic says quiet morning on low-pressure seeds. That might be exactly backwards for a first run. + +--- + +## Questions for Jeroen + +### Q1: Did the game feel like it had anything at stake in the first 8 minutes? + +Not "did you understand the stakes intellectually" — you designed the game, you know what's at stake. But emotionally, in the moment: did anything feel like it *mattered*? Did you care what happened to anyone? Was there anything you wanted to *protect* or *find out* before you hit the testing wall? + +**Why this matters:** The seed variants and dual-lens divergence are replay engines built on top of emotional investment. If minute 1-8 produces zero investment — no person you care about, no situation that pulls you forward, no question you want answered — then the replayability architecture is a cathedral with no foundation. The structural variety is real. But variety of *what*? If the base experience produces no attachment, replaying it produces varied nothing. + +--- + +### Q2: The "no objectives, no markers" philosophy is theoretically great for replayability — checklist games are solvable, emergent games aren't. But in practice, did "no objectives" feel like freedom or abandonment? + +Specifically: if you had played this as a new player with no design knowledge, do you think you would have formed a self-generated goal — "I want to find out what Kael is hiding," "I want to understand this district" — by minute 15? Or would you have closed the game before that goal could form? + +**Why this matters:** The no-objectives design creates replayability by preventing the game from being "solved" via checklist. But it requires the player to stay long enough to generate their own investment. If the window to generate that investment is longer than the window before a new player quits, the philosophy defeats itself. I need to know whether the current design can produce a self-generated goal in time, or whether we need a first-run scaffolding layer that doesn't compromise the no-tutorial philosophy for repeat players. + +--- + +### Q3: The storyteller selects "quiet morning" (Variant A) on low-pressure seeds. Should the first playthrough get a different storyteller configuration than replays? + +The quiet morning variant is designed to make the crack feel retroactively obvious on a second playthrough — you look back and think "it was right there." That's a second-playthrough payoff. On a first playthrough, "completely normal day" with no visible anomaly until minute 10+ might be the exact wrong opening for a player who doesn't yet know what they're watching for. + +Should the storyteller detect "first playthrough" and deliberately select a higher-pressure seed — Kael late, container held — where the anomaly appears faster, even if that means the "quiet morning" retroactive punch is lost? Or does giving the first-run player a stronger signal betray the design? + +**Why this matters:** If "first playthrough" and "optimal replay structure" require different storyteller behavior, that's actually a good sign — it means the game has enough depth to need it. But it also means we're currently deploying the replay-optimized configuration on players who haven't yet earned it. The replayability EXPLODES if people get to playthrough 2. The question is whether the current first-run experience gets them there. diff --git a/docs/workshops/wheres-the-fun/round1-ozzie.md b/docs/workshops/wheres-the-fun/round1-ozzie.md new file mode 100644 index 000000000..9141dabc7 --- /dev/null +++ b/docs/workshops/wheres-the-fun/round1-ozzie.md @@ -0,0 +1,58 @@ +# Round 1 — Ozzie: Diagnosis Questions + +**Workshop:** Where's the Fun? v0.1 Playtest Reckoning +**Agent:** OZZIE (Player Experience & Wow Factor) +**Round:** 1 — Diagnosis + +--- + +## What I'm Looking At + +The wow moments checklist is brutal. 1 of 23 content deliverables ready. All 6 moments: "Not started." The monologue system (#119, #120, #122) is in backlog. Observation event generator (#239): backlog. THE FRIEND content: doesn't exist. + +The wow moments didn't land because THEY WEREN'T THERE. The playtest ran on UI stubs, fog, and movement. No authored lines. No anomaly detection. No chime. No friend. No revelation. + +That's actually two separate problems: + +1. The content pipeline is 5% built — that's a resourcing and sequencing problem. +2. The player hit a wall at minute 8 and STILL couldn't figure out what to do — that's a design problem that exists INDEPENDENT of content completion. + +I need to know which one Jeroen felt more. + +--- + +## Questions for Jeroen + +### Question 1: The Gut Feeling + +**"Close your eyes. The Settled Reach is finished and it's genuinely fun. What are you doing in that moment? Not the design doc version — your gut. What does your body respond to?"** + +*Why this matters:* Every design decision in this project traces back to a vision of the fun. But the playtest broke the feedback loop. Before we diagnose what went wrong, I need to know if the vision itself is still vivid and real to Jeroen, or if the playtest knocked it loose. If Jeroen can describe a clear, visceral moment — "I'm watching an NPC break their routine and my character says something that makes my stomach drop" — that's a target worth rebuilding toward. If the answer is vague or hesitant, the design intent may be more fragile than we thought. + +--- + +### Question 2: The Ghost Question + +**"The checklist shows 1 of 23 wow-moment deliverables is ready. Those moments weren't in the build. When you hit the wall at minute 8 — was your frustration 'I don't have enough yet' or something scarier: that even if all 6 moments existed tomorrow, you still wouldn't have known to CARE?"** + +*Why this matters:* There are two very different failure modes here. Failure Mode A: great design, not yet built. The fix is resourcing — write the content, implement the pipeline, run the playtest again. Failure Mode B: the moments are beautiful on paper but the player has no reason to be emotionally invested when they arrive. THE FRIEND's Contradiction requires 20 minutes of attachment before the payoff. The Character's Eye requires the player to trust that the monologue means something. If the player doesn't care about the character at minute 5, the minute-20 revelation doesn't land no matter how well it's written. + +Which failure mode did the wall feel like? + +--- + +### Question 3: The Floor Before the Ceiling + +**"The 6 wow moments are all minute 5-25 payoffs. The testing wall hit at minute 8. Is there a minute-0-to-5 problem that is completely separate from those moments — something about the first 60 seconds that has nothing to do with content completion?"** + +*Why this matters:* The wow moments are the ceiling. But something is wrong with the floor. The player stepped into the world and within 8 minutes gave up — not because the content wasn't ready, but because there was no pull forward. No itch. No "I wonder if...". In Hamilton's universe, characters who walk into rooms feel IMMEDIATE purpose even when they're lost. Did the player feel like they were a person arriving somewhere? Or did they feel like they were operating a camera with no subject? That first-60-seconds feeling — intentional or broken — is the foundation everything else stands on. If we can't fix the floor, the ceiling doesn't matter. + +--- + +## What I'm Listening For + +- Whether the design vision is still emotionally alive for Jeroen, or whether the playtest hollowed it out +- Whether the gap is "not built yet" vs "wrong design" +- Whether there's a pre-wow-moment problem — a minute-0 hook — that doesn't exist anywhere in the current design docs + +THAT'S where the fun lives or dies. diff --git a/docs/workshops/wheres-the-fun/round1-paula.md b/docs/workshops/wheres-the-fun/round1-paula.md new file mode 100644 index 000000000..7d7287375 --- /dev/null +++ b/docs/workshops/wheres-the-fun/round1-paula.md @@ -0,0 +1,60 @@ +# Round 1: Paula — Diagnosis Questions +## Where's the Fun? Workshop | 2026-03-05 + +--- + +## My Frame + +The smuggler's moral arc is the emotional engine of the v0.1 vertical slice. Four phases, authored FactId gates, irreversible transitions. The design is sound on paper: start comfortable, get pulled into doubt by human cost, face reckoning, carry the weight. The arc depends entirely on the player forming *attachment* to specific people before the system delivers moral consequence through them. + +The playtest broke at minute 8. Which means the player never reached Phase 2. Which means the arc never started. + +The honest truth is: this is not a content problem. You can't have a moral arc about people you haven't met. If Naia, Maret, and Pell are background tiles rather than people with stakes, the Phase 1-to-2 gates never fire — not because the system is broken, but because the player has no reason to watch them closely enough to notice their distress. + +Three questions probe this. + +--- + +## Question 1: Did Phase 1 feel like a *home* before it felt like a threat? + +**The design dependency:** The smuggler's arc depends on Phase 1 (Comfort) feeling *good* — operationally confident, fond of Kael, the operation manageable, nobody visibly hurt. Phase 2 lands as loss of that comfort. Phase 3 lands as reckoning against it. If Phase 1 never felt like warmth, there's nothing to lose. + +**The playtest suggests:** You hit a wall before knowing what to do or why to care. That's Phase 1 failing as a home. The smuggler's baseline comfort should provide implicit purpose — *I have an operation to run, these are my people, this place is familiar* — without needing to be instructed. + +**The question:** In the minutes you played, did the smuggler's starting relationship to Kael, to the dock, to the operation, feel like a *stable world you belonged to* — even briefly? Or did it feel like a stranger in an unknown place with unknown tasks? + +**Why it matters:** If Phase 1 never felt comfortable, the entire arc is rootless. The fix is different depending on the answer. If it's a content problem (Phase 1 lines aren't warm enough), that's one fix. If it's a mechanical problem (no behavioral anchor to the operation), that's another. + +--- + +## Question 2: Did any NPC feel like a *person with stakes* before you stopped playing? + +**The design dependency:** The Phase 1-to-2 transition gates on observing human cost — Naia's stress at the bar, Maret double-checking manifests, Pell's voice shaking. These are not dramatic reveals. They're quiet behavioral tells. The player has to be watching the right person at the right time AND have enough prior context to read the tell as *significant* rather than decorative. + +**The playtest suggests:** "Monologue feels scattered and contextless" (issue #2) and "monologue observations disconnected from visual source" (issue #9). If a chalk-mark observation fires without the player knowing what chalk marks mean, it's noise. The same is true for moral observation lines — if the player doesn't know who Naia is, a monologue line about Naia looking tired at the bar is indistinguishable from ambient flavor. + +**The question:** By the time you stopped playing, had any single NPC — Kael, Naia, Maret, Pell, anyone — registered as someone whose state *mattered to you personally*, not just as an object in the simulation? Did anyone feel like a *person* rather than a moving tile? + +**Why it matters:** But what SUSTAINS the moral arc across 30 minutes is specific attachment to specific people. If the answer is "no," the problem isn't the arc's phases — it's the absence of relationship establishment as a designed beat before the arc can function. The arc needs a precondition that doesn't currently exist in the opening sequence. + +--- + +## Question 3: Did you feel *complicity* while playing, or only recognize it as a design intent? + +**The design dependency:** D-091 establishes complicity as the thematic core — not a reward, not a twist, but the slow realization that you've been part of something from the start. The smuggler's arc is built around the moment this becomes undeniable. The design goal is felt experience, not intellectual recognition. + +**The playtest suggests:** You designed this game. You know the Kael contradiction is coming. You know the moral weight of every manifest filed. You know what "innocent" means as a `moral_weight` value. A naive player has none of this meta-knowledge — and from the playtest evidence, they also have no in-world signal that the smuggler's operation has any moral dimension at all. It's just movement and fog. + +**The question:** Let me complicate this by separating two things: did you feel *any* moral pull — guilt, unease, protective instinct, the sense that someone might get hurt — *while playing*? Or did the complicity only exist at the design level, as something you knew was there but couldn't feel through the current implementation? + +**Why it matters:** If complicity is currently only intellectual (you know it's the theme), the arc may need an earlier, cruder, more direct trigger — something that forces the player to make a small choice with visible consequences *before* they've formed the vocabulary to understand the moral arc. The 4-phase authored arc is the right destination. The question is whether there needs to be a simpler on-ramp that produces *felt unease* in the first 3 minutes, before the FactId gates can do their work. + +--- + +## What I'm Listening For + +The diagnostic I'm building toward: the moral arc is structurally sound but *temporally displaced*. It's designed for a player who is already attached, already oriented, already at home in Phase 1. The playtest failure suggests that player never exists in v0.1 — the arc starts in a void. + +If Jeroen confirms that no NPC registered as a person with stakes, and no complicity was felt (only recognized), the recommendation will not be "revise the arc." It will be "add a relationship establishment phase before the arc begins" — a designed beat that makes the smuggler's warmth toward Kael and the normalcy of the operation *felt*, not assumed. + +The arc doesn't need to be simpler. It needs an earlier floor. diff --git a/docs/workshops/wheres-the-fun/round1-tyre.md b/docs/workshops/wheres-the-fun/round1-tyre.md new file mode 100644 index 000000000..00243a7c7 --- /dev/null +++ b/docs/workshops/wheres-the-fun/round1-tyre.md @@ -0,0 +1,72 @@ +# Round 1: Tyre — Diagnosis Questions + +**Role:** Technical Architect +**Sources reviewed:** Workshop brief, `decisions/architecture.md` (D-010, D-020, D-026, D-041), `docs/design/interaction-verbs-v0.1.md` + +--- + +## Technical Context (for the other agents) + +*cracks knuckles* — Before the questions, a quick reality-check on what the architecture can and can't absorb. This matters for scoping proposals in Round 3. + +**What's cheap to change (server-only, days not weeks):** +- Monologue priority, timing, cooldown, queue depth — all tuning parameters in the server's monologue system. Rebalancing the signal-to-noise ratio (playtest item #10) is a config change, not an architecture change. +- Adding new fields to `ObserverSnapshot` — the MessagePack protocol is variable-shape by design (D-020). The client renders what it receives. Adding, say, a "current objective hint" or "knowledge progress summary" to the snapshot is a server-side addition + a client-side UI widget. No protocol rewrite. +- Verb priority shifts — the `verbs[]` array is already computed server-side every tick with full context (distance, relationship state, knowledge level). Changing what `[E]` does in different situations is priority reordering, not new systems. +- Knowledge graph queries — the graph already tracks what the player knows, what's contradicted, what's stale. Any "progress" system can be built as a read-only query over existing data. The data is there; we just don't surface it. + +**What's moderate (new systems, but within existing architecture):** +- An objective/hint system that reads the knowledge graph and emits guidance through the existing ObserverSnapshot pipeline. This is a new server-side system, but it plugs into existing infrastructure. Think: a system that runs once per game-minute, checks knowledge state, and pushes a "current thread" indicator to the client. Maybe a week of work. +- Spatial anchoring for monologue (playtest item #9) — tying monologue triggers to specific tile positions so the client can render an indicator at the source location. Requires adding a `source_position` field to the monologue payload. Server change + client rendering change. + +**What's expensive (architecture-level, would delay v0.1):** +- Branching dialogue trees replacing the tagged line pool system (D-028). The entire content pipeline assumes pool-based selection. This would be a rebuild. +- Real-time NPC conversation AI (dynamic dialogue generation). Not in the architecture at all. +- Fundamentally changing the client-server boundary (moving logic to the client). This contradicts D-010/D-020/D-048 and would be a project reset. + +--- + +## Question 1: The Knowledge Graph Already Knows — Why Doesn't the Player? + +The knowledge graph (D-041) tracks everything the player character has learned: entity knowledge, fact knowledge, confidence levels, contradictions, staleness. The server computes this every tick. But none of it is surfaced to the player as *structured progress*. + +The monologue system was supposed to bridge this gap (D-016), but it fires as prose fragments — atmosphere, not information architecture. The player gets "He checked his lattice again. Third time." but never gets the structured signal: *"You've noticed 2 of 3 tells on this NPC. Something is off here."* + +**The question:** When you designed monologue as the primary feedback mechanism, were you envisioning it as the *only* channel between the knowledge graph and the player? Or was there always an implicit assumption that some structured feedback layer (a journal, a "threads" panel, a knowledge summary) would eventually sit alongside it? + +**Why this matters technically:** If the answer is "monologue was supposed to be sufficient," then the fix is content and tuning — write better monologue lines, tighten the priority system, add spatial anchoring. If the answer is "there was always supposed to be more," then we need to design that structured layer now, and the good news is the architecture supports it cheaply — the data already exists in the knowledge graph, we just need a new rendering path through ObserverSnapshot to a client-side UI element. + +--- + +## Question 2: What Does "No Objectives" Actually Mean at the Architecture Level? + +The playtest showed that "no objectives, no markers" produces paralysis, not discovery (item #13). But "objectives" is a spectrum, not a binary. Technically, the architecture can support anything from: + +- **Tier 0 (current):** Nothing. The player infers purpose from monologue and observation. +- **Tier 1:** Passive knowledge summary — a panel showing what the player character knows, organized by entity/fact, with contradictions highlighted. No direction, just structured memory. *"You know these things. Some of them conflict."* +- **Tier 2:** Diegetic threads — the character's neural insert (already established in D-013) surfaces "active threads" based on knowledge graph state. *"Kael's schedule doesn't match what Torek said."* Not objectives — observations the character is actively puzzling over. +- **Tier 3:** Explicit objectives — a mission briefing, checkpoints, "investigate X." Traditional game objectives dressed in diegetic clothing. + +All four tiers are technically feasible within the existing architecture. Tiers 1-2 are moderate effort (query existing knowledge graph, new UI widget). Tier 3 is also moderate but requires authored objective content. + +**The question:** Where on this spectrum does the game need to land for v0.1? And critically — is the resistance to objectives a *design principle* (the game is fundamentally about discovering your own purpose) or a *scope decision* (we didn't build it yet)? Because architecturally, I can support any of these without touching the simulation core. + +**Why this matters:** If it's a design principle, we're solving the fun problem through better monologue, spatial feedback, and environmental storytelling — and the architecture is fine as-is. If it's a scope decision, then the cheapest high-impact change might be a Tier 2 "threads" system — maybe 3-5 days of server work, 2-3 days of client UI — that gives the player structured awareness of what their character is tracking, without ever saying "go here, do this." + +--- + +## Question 3: Is the Single Context Key Hiding the Game From the Player? + +The v0.1 verb system has 7 verbs, but the player only sees one at a time through `[E]`. The system auto-selects based on priority. The player presses `[E]` and gets whatever the server decided was most important. The design intent was simplicity — don't overwhelm with choices. + +But the playtest suggests the opposite problem: the player doesn't know what actions *exist*. They don't know they can Examine NPCs separately from Talking. They don't know Overhear is a thing that happens. They don't know monologue is reactive to their observations. The verb system is invisible. + +The v0.2 plan already calls for showing the full `verbs[]` array (e.g., `[E] Talk [F] Observe`). This requires **zero server changes** — the protocol already carries all available verbs. It's purely a client-side rendering change. + +**The question:** Would pulling the v0.2 verb display forward to v0.1 — showing all available verbs instead of just the top one — help solve the "I don't know what to do" problem? Or would it just add more noise to an already overwhelming experience (item #10)? + +**Why this matters:** This is the single lowest-effort change that could materially affect the core loop. It's 1-2 days of client work, zero server work. If the player can *see* that "Observe" exists as a distinct action from "Talk," they might naturally discover the investigation loop. But it only helps if the player's problem is "I don't know my options" rather than "I don't know why I should care." If it's the latter, showing more verbs just adds clutter. + +--- + +*Feasibility summary: the architecture is more flexible than the playtest suggests. The simulation engine doesn't need rebuilding — it needs better channels between what it already knows and what the player can see. The knowledge graph is rich; the rendering of that knowledge to the player is impoverished. Most high-impact changes are in the "moderate" category: new server-side query systems feeding new client-side UI, all within existing protocol boundaries.* diff --git a/docs/workshops/wheres-the-fun/round3-araminta.md b/docs/workshops/wheres-the-fun/round3-araminta.md new file mode 100644 index 000000000..d881ad77f --- /dev/null +++ b/docs/workshops/wheres-the-fun/round3-araminta.md @@ -0,0 +1,125 @@ +# Round 3: Araminta — Visual Design Proposals +## Where's the Fun? Workshop | 2026-03-05 + +--- + +## The Reframe, From Where I Sit + +The interview confirmed the diagnosis I was hoping was wrong: both reading AND acting were failing simultaneously (Q5). Visual hierarchy "would have helped" (Q17) — but the deeper finding is more alarming than a hierarchy problem. NPCs did not register as people at all (Q6, Q7). Sova Transit was "game level with dots" (Q14). Zero legibility. Zero attachment. Zero foundation for any emotional, moral, or mechanical loop to run on. + +The life-sim reframe (Q1, Q15) makes this problem bigger, not smaller. A life sim is entirely load-bearing on the readability of people and place. If you can't tell who someone is, what they do, and what kind of space you're in — you cannot live in the world, build relationships, build businesses, or have impact. The social simulation is invisible because the social substrate has no visual form. + +This is not a polish problem. Character and place legibility are the structural foundation of the game. Everything else — the moral arc, the asymmetric information mechanic, the knowledge graph, the monologue — assumes the player can SEE who and where they are. + +--- + +## KEEP + +**The 3-tier visual hierarchy concept.** Jeroen confirmed it would have helped even in the broken v0.1 state. Keep this as the organizing rule for ALL feedback signals: danger (critical, immediate), relevant (actionable, player-adjacent), ambient (world texture, not requiring action). This carries directly into v0.2 and becomes more important as the feedback tool suite expands. + +**The diegetic UI philosophy.** Inserts, overlays, in-world anchors. No floating HP bars. Nothing that breaks the fiction of being a person in a place. Jeroen's vision of job-specific information tools (mystery board, journal, AR overlays) confirms this direction — the player's HUD is their neural insert, and different careers have different inserts. Keep this as an absolute rule: every UI element belongs either to the world or to the character's insert system. + +**Fog and perception rendering.** The technical pipeline works. Keep it. Fix the fog edge (item 3 — pure visual bug, no mechanical dimension), then leave the system alone. It's not the problem. + +**The principle of contextual name reveal.** Even in the current implementation, interaction prompts obfuscate NPC identity until resolved (design intent from item 7). Keep this. It's the right visual mechanic for a knowledge-based game — the player's visual information tracks their character's knowledge. A stranger is "Unknown Person" until you have a reason to know more. + +--- + +## CHANGE + +### 1. Character Legibility — Dots Must Become People (Priority 1) + +This is the single most important visual design problem in the project. Not because it's visually broken, but because nothing else can function without it. + +**The problem:** All NPCs are visually equivalent dots. The player cannot tell dock workers from Commission officers from bar staff from suspects. There's no visual hook to begin forming identity, tracking individuals, or reading behavioral state. + +**The strategy — three layers:** + +**Layer 1: Archetype silhouettes and color.** Each NPC archetype gets a distinct body-language posture and a palette anchor. This doesn't require full character art — even at the box-and-label stage, dock workers should carry themselves differently than officials. Color is the fast read: Commission blue, dock-worker orange, civilian grey, bar staff warm amber. These map to faction and role simultaneously. + +**Layer 2: Behavioral state reads.** NPCs in routine comfort vs stress vs suspicion should be visually distinguishable at distance without labels. At the box-and-label level: simple icon indicators above sprites (small, low-weight — ambient tier). Stress shows as a muted amber pulse. Suspicion as a directional indicator. These disappear as art matures, replaced by actual animation and body language. The visual grammar exists from day one. + +**Layer 3: Knowledge-gated reveal.** What the player SEES about a person tracks what their character KNOWS. Unknown person: silhouette, archetype color, no name. Observed (in view >10s): role label appears ("Dock Worker"). Identified (spoken to or monologue trigger): name appears. Investigated (knowledge graph entry): status indicators become readable. This is the same knowledge-graph architecture expressed visually. The screen IS the knowledge graph. + +### 2. Place Identity — Tile Grid Must Become an Inhabited Space (Priority 1, parallel) + +**The problem:** Sova Transit reads as "game level." No functional cluster is visually distinct. The player cannot tell bar from dock from corridor from supervisory area. + +**The strategy:** + +**Functional cluster palette system.** Each cluster type has a distinct base palette that the player will learn to read before they can articulate why. Bar: warm, slightly worn, amber tones. Dock logistics: cool industrial, blue-grey metal. Corridors: transitional neutral. Supervisory/administrative: cleaner, lighter, institutional. Even with placeholder tiles, the COLOR TEMPERATURE of the floor tiles communicates the function of the space. + +**Ambient life props.** Crates, drink containers, work documents, seating — present in the tile layer as visual noise that reads as "people use this space." Not gameplay-interactive (yet), but visually communicating that the space has been inhabited. This is cheap and high-impact: a single prop palette per cluster type costs one sprint and changes the reading of the space dramatically. + +**Depth and layering.** Top-down readability improves dramatically with foreground/background separation. Even placeholder boxes can be layered: equipment in the background plane, characters in the mid-plane, immediate environmental details in the foreground. The player gains spatial orientation from layering that flat sprites don't provide. + +### 3. HUD Visual Hierarchy — All Signals Through the 3-Tier System + +**The problem:** Item 10 — everything has equal visual weight. The player cannot triage. Five simultaneous signals means zero actionable signal. + +**The change:** Every HUD element, every monologue line, every sound indicator, every behavioral cue gets assigned a tier before it's implemented. No exceptions. + +- **Danger tier** (critical red, high contrast, animated): immediate threat, time pressure, consequence now. Rare. Use sparingly or it loses meaning. +- **Relevant tier** (amber, medium weight, static): actionable by the player right now. This is the tier the player should be watching. +- **Ambient tier** (muted grey/white, low weight, brief fade): world texture, present but not demanding attention. Most signals live here. + +This is a design rule, not just a visual rule. Content authors need to know which tier their signal belongs to BEFORE it gets written and implemented. The visual designer's job is to enforce the tier system as a gating criterion — if it's not assigned a tier, it doesn't ship. + +### 4. UI Chrome and Anchoring + +**Context menu (item 8):** Needs panel chrome — a semi-transparent, insert-styled frame that communicates "this is your neural insert presenting options." Not a floating text list, not a generic dialog box. The visual language says: you are seeing through your character's perception system. Design: rounded-corner dark panel, insert accent color, option text weighted heavier than labels. + +**Stance indicator (item 4):** Remove from free-float. Anchor to the minimap frame or the insert HUD band. Icon + label + keybind, always visible when a stance is active. The stance is a persistent state — it needs a persistent home, not a tooltip. + +**Monologue spatial anchor:** Text appears in a consistent panel zone, but the SOURCE of the observation should have a visual cue in the world — a brief, low-weight highlight on the tile or entity that triggered the thought. This addresses item 9 (observations disconnected from visual source) without requiring complex systems. A 0.5s highlight at ambient tier cost is enough. + +--- + +## KILL + +**"Visual polish is low priority" (item 6).** This framing is wrong and needs to be retired. In a game where visual legibility IS the primary mechanic — where the player reads the world to understand their information state — there is no visual element that's decorative-only. The floor of screen readability is load-bearing. Polish conversations can prioritize specific items, but no visual element should be tagged as categorically low priority. + +**Floating, unanchored UI elements.** Every UI element needs a home in either the world space (diegetic, attached to an entity or location) or the insert space (the player character's neural overlay, in screen-space with consistent positioning rules). Elements that float freely between these two frames break the visual grammar and undermine trust in the interface. Kill the pattern, not just the specific instances. + +**Uniform NPC appearance.** Every dot looks the same. This is the most direct visual cause of the NPCs-as-dots problem. Kill this immediately — even before full sprite work, archetypes need palette differentiation. One sprint, massive readability gain. + +--- + +## The Career Insert Visual Strategy (Carries Into v0.2) + +The life-sim reframe with CK3-style bookmarks (Q15) creates a direct visual design opportunity: each career path has a distinct visual identity for their insert HUD. + +**Law enforcement insert:** Clean institutional design. Commission blue accent. Behavioral flag overlays (suspect status, known associations). Information presented as case-file aesthetic — labeled, formal, authoritative. + +**Tycoon insert:** Financial overlay. Asset markers on the map. Investment opportunity indicators. Warmer palette, market-board aesthetic. + +**Smuggler insert:** Street-contacts network. Social graph edges visible between known NPCs. Risk indicators on patrol patterns. Grittier, analog-feeling — like a handwritten ledger overlaid on the world. + +Each insert teaches the player what KIND of person they are through the visual language of their information system. This is the "no objectives, just diegetic tools" philosophy made concrete. The insert IS the HUD IS the character sheet. + +**Design rule:** The insert visual grammar is defined ONCE, in the style guide, before any individual insert elements are implemented. Otherwise each feature team invents their own overlay aesthetic and the coherence collapses. + +--- + +## What Carries Directly Into v0.2 + +1. **3-tier signal hierarchy** — implemented as a design gate, not optional +2. **Character archetype visual system** — silhouettes, palette anchors, behavioral state indicators +3. **Knowledge-gated reveal progression** — visual state tracks knowledge state +4. **Functional cluster palette system** — place legibility through color temperature +5. **Career insert visual grammar** — one design document, defines the HUD language for all career paths +6. **Diegetic anchor rule** — every UI element belongs to world-space or insert-space, no exceptions + +--- + +## The Priority Order + +1. **Character legibility** — archetype palette + behavioral state indicators. One sprint. Prerequisite for everything else. +2. **Place legibility** — functional cluster palettes + ambient props. One sprint, parallel to character work. +3. **HUD hierarchy enforcement** — retroactively assign tiers to all existing signals, fix the outliers. Two-week sweep. +4. **UI anchoring** — context menu chrome, stance indicator home, monologue spatial anchor. One sprint. +5. **Career insert grammar** — define the visual language for all bookmark paths. Design-only sprint, no implementation. + +None of these require full art. All of them work at the box-and-label fidelity level. And all of them are prerequisites for evaluating whether the life-sim loop is working — because right now, the playtest can't tell us anything useful about whether the game is fun. It can only tell us the player can't read the screen. + +Fix the screen first. Then ask if the game is fun. diff --git a/docs/workshops/wheres-the-fun/round3-gestalt.md b/docs/workshops/wheres-the-fun/round3-gestalt.md new file mode 100644 index 000000000..095251cbc --- /dev/null +++ b/docs/workshops/wheres-the-fun/round3-gestalt.md @@ -0,0 +1,192 @@ +# Round 3: GESTALT — Systems Design Proposals +## Workshop: Where's the Fun? | 2026-03-05 + +**Agent:** GESTALT (Systems Design) +**Based on:** Full interview transcript, Round 1 notes, all 9 Round 1 files + +--- + +## The Reframe in Systems Terms + +Jeroen's vision is: *live in the world, have impact*. Single-character life sim. Detective is a job you might have, not the whole game. This isn't a tweak to the design — it's a complete recontextualization of what every system is FOR. + +The good news mechanically: the simulation engine (routines, perception, knowledge graph, verb architecture) was already built for this. The bad news: the v0.1 scope stripped the player of all their planned interfaces to that engine and then asked them to play. The playtest failure was a scaffolding problem, not a simulation problem. + +Let me break down what this means for every system I own. + +--- + +## KEEP — Working, Needs Recontextualization + +### 1. The verb architecture (`verbs[]` system) + +The server-computes-all / client-reads-priority design is architecturally correct and survives the reframe. The fact that every entity gets a full `verbs[]` array computed every tick, and the client just reads `verbs[0]`, is exactly the right foundation for a life-sim where different jobs need different default actions. + +Keep the architecture. The priority profiles need to become job-aware (see CHANGE section). + +### 2. The knowledge graph + +It works. The data exists. Jeroen confirmed the planned full information architecture (mystery board, journal, AR overlays, comms) was always assumed — the knowledge graph is the right backend for all of it. Don't touch it. + +### 3. The perception system (fog, LOS, shadowcasting) + +Fires correctly. The life-sim reframe needs this even more than the detective game did — your spatial position determining what you know is MORE interesting when what you're doing with that knowledge varies by job. + +### 4. The asymmetric information mechanic itself + +This becomes MORE powerful in a life-sim, not less. In the detective-only framing, asymmetric information meant "the detective and smuggler know different things." In the life-sim framing, it means "every job gives you a different lens on the same world." A detective in Sova Transit notices behavioral anomalies. A tycoon notices cargo flow. A smuggler notices authority figures. Same tiles, same simulation, three different games. This is the correct master mechanic. It just needed the broader frame to breathe. + +### 5. The moral arc patterns + +The 4-phase arc (belonging → crack → investigation → reckoning) survives as world content — specifically, as what happens when a player pursuing ANY career path encounters the FRIEND/smuggler storyline. It's not the core loop anymore; it's one of the world's live storylines that may or may not intersect with the player's current job. Keep the patterns. Release them from being the mandatory structure. + +--- + +## CHANGE — Design Intent Right, Execution Wrong + +### 1. Verb priority must become job-aware + +**Current spec:** Talk > Examine NPC at close range (default state) + +**The problem:** This was designed for a detective game where the primary loop was observation-then-confrontation. But even within the detective game, it contradicted the stated loop ("Examine first, Talk later"). In the life-sim reframe, different jobs need different default interactions — a tycoon wants to Examine Objects (read manifests, terminals, ledgers); a bar owner wants to Talk; a detective wants to Observe. + +**The change:** Introduce `VerbPriorityProfile` as a property of the player's current career/job. The server already has all the context to compute this — it knows the player's job, their relationship state, their knowledge graph. The `verbs[]` computation adds one more input: active `VerbPriorityProfile`. + +``` +VerbPriorityProfile { + detective: [ExamineNpc, Talk, ExamineObject] // Observe first + tycoon: [ExamineObject, Talk, ExamineNpc] // Read the room first + smuggler: [Talk, ExamineNpc, ExamineObject] // Social first + bar_owner: [Talk, ExamineNpc, ExamineObject] // Same as smuggler +} +``` + +This is a moderate server change — a new field on the character entity, a refactor of the priority sort step. No client changes needed beyond rendering whatever `verbs[0]` says. + +### 2. Monologue demoted from primary to supplementary + +**Current spec:** Monologue is the primary feedback mechanism bridging the top-down camera to the character's subjective experience (D-016). + +**What Jeroen confirmed:** Monologue was always meant as "a reasoning nudge and summary tool in the later stages." It was supposed to supplement a full diegetic tool suite, not carry the entire feedback burden alone. + +**The change:** Monologue stays — but its design role must be explicitly restated. It is NOT the tutorial, NOT the primary feedback channel, NOT the player's only window into the knowledge graph. It is: internal color, atmospheric texture, reasoning confirmation ("yes, you did notice something real"). The job onboarding is the tutorial. The diegetic tools are the information architecture. Monologue is the voice that runs on top of both. + +Operationally, this means: +- Fewer monologue lines overall (not the only thing firing) +- Higher bar per line (if it fires, it earns its moment) +- Spatial anchoring (source position) becomes a requirement, not a stretch goal +- Monologue queue depth can stay at 1 — but the 2s cooldown may need loosening since the player now has other feedback channels carrying load + +### 3. The storyteller: from seed variant selector to career onboarding engine + +**Current design:** Storyteller picks from detective/smuggler variants (quiet morning, Kael late, container held, etc.) + +**What Jeroen described:** CK3-style bookmarks. Choose a career path. Each career path has onboarding that teaches tools organically. Detective starts with job onboarding (insert enablement, weapon qualification), not a mission. Tycoon starts with the inheritance ping. + +**The change:** The storyteller's primary job becomes: +1. **Bookmark resolution** — which career path did the player choose? +2. **World state initialization** — what is the day's world state? (tensions, schedules, active storylines) +3. **Onboarding sequencing** — which tools get enabled in which order during the first session? + +The seed variants (quiet morning, Kael late) survive as sub-options within career paths — they can still modulate the world state after onboarding is complete. But they are no longer the FIRST thing the storyteller decides. Career + onboarding comes first. + +### 4. The "not a tutorial" philosophy — valid principle, wrong implementation + +The principle was right: the game teaches through play, not instruction. The implementation was wrong: it assumed the player would naturally discover the verbs and loops through exposure alone, with monologue as the only scaffold. + +In the life-sim reframe, the "not a tutorial" principle is preserved through job onboarding. Each career bookmark's onboarding is the tutorial — but it's diegetic: +- Detective: "Your insert has come online. New capability: Pattern Tracking. High-probability behavior anomalies will surface automatically." The player learns Pattern Tracking exists because the insert tells them their own character's professional context. +- Tycoon: "Inheritance received. Portfolio summary available via insert." Player learns portfolio exists because the bank contact tells them. + +The tutorial is still not a pop-up. It's the world talking to you in-character. The difference from v0.1 is that the world now has something to say that isn't just atmospheric monologue. + +--- + +## KILL — Design Intent Was Wrong + +### 1. The monologue-only feedback architecture + +Kill it as a design intent. Not the monologue system — that survives. But the architectural decision that monologue is sufficient as the player's primary interface to the knowledge graph is wrong. It was always wrong; Jeroen confirms this was a scoping error. The planned tools (mystery board, journal, AR overlays, comms, insert icons) should be treated as CORE, not aspirational. + +For v0.2 design, these tools need to be treated as first-class citizens: their absence in v0.1 WAS the playtest failure. Not bad monologue lines — missing interfaces. + +### 2. The single detective/smuggler frame as the game's identity + +Kill the framing that the game is fundamentally a detective/smuggler tension. That storyline lives in the world. It may still be the most interesting storyline in the world. But the game is not defined by it. The v0.1 scope was too narrow. In v0.2, the frame is: you are a person in the Settled Reach. What do you do? + +This has downstream implications for every design doc that frames every decision as "for the detective" or "for the smuggler." Those need to be reframed as job-specific instances of broader patterns. + +### 3. The verb priority spec as written (Section 4, interaction-verbs-v0.1.md) + +The specific priority table is wrong and needs to be replaced with the `VerbPriorityProfile` approach described above. The section header "v0.1 Priority Resolution" is particularly misleading — it implies the priority is global and static, when it should always have been job-contextual. + +--- + +## Where Does the Fun Come From? + +*The positive vision for systems design in the life-sim.* + +**The fun is in reading the same world differently depending on who you are.** + +Every system I own is in service of this. The perception system, knowledge graph, verb architecture, and storyteller all exist to create: the experience of having a specific perspective on a shared world. Not one perspective. Many. Simultaneously incompatible. + +The mechanic that produces this isn't "asymmetric information about the detective case." It's "asymmetric information about everything, filtered through your job." + +Here's what this looks like as a systems interaction: + +``` +Player character (Detective) enters cargo bay +↓ +Perception system: LOS reveals 3 NPCs, 2 cargo terminals, 1 supervisor +↓ +VerbPriorityProfile (detective): default = Observe NPCs +↓ +Knowledge graph query: any known contradictions on these entities? +↓ +Insert UI: flags Kael's behavioral anomaly count (2/3 tells observed) +↓ +Monologue (if appropriate): "Third time this shift. He keeps checking the gate log." +↓ +Player decision: approach (Observe or Talk), examine terminal, wait and watch +``` + +The same cargo bay, entered by a tycoon: + +``` +Player character (Tycoon) enters cargo bay +↓ +Perception system: same LOS, same NPCs, same terminals +↓ +VerbPriorityProfile (tycoon): default = Examine Objects +↓ +Knowledge graph query: any cargo flow anomalies? +↓ +Insert UI: flags the held container (3 days over routing protocol) +↓ +Monologue: "Someone's paying demurrage on that. Either they can afford it or they can't afford not to." +↓ +Player decision: examine terminal, talk to supervisor, calculate the margin +``` + +Same simulation. Same world state. Different games. THAT is the asymmetric information mechanic. THAT is where the fun is. + +The systems I need to build to make this real: +1. `VerbPriorityProfile` per career/job (moderate server work) +2. Job-specific insert UI layouts (moderate client work) +3. Knowledge graph surface queries per job (what does THIS job flag as salient?) +4. Career bookmark selector + onboarding arc sequencer (new storyteller subsystem) +5. Diegetic tool suite as first-class interfaces, not stretch goals + +--- + +## Specific Tensions to Resolve in Round 4 + +1. **Career vs. class arc tension** — Jeroen wants both job-specific onboarding AND storylines that cut across jobs. How does the gate-builder/smuggler storyline intersect with a tycoon playthrough? The moral arc needs to be job-agnostic (it's about WHAT YOU KNOW, not WHAT YOUR JOB IS). This needs cross-domain discussion. + +2. **NPC legibility is a prerequisite for everything else** — Araminta is correct that visual hierarchy comes first. But it's not just visual. The NPCs need to be legible as PEOPLE before the knowledge graph can produce interesting decisions about them. This requires art, names, behavioral tell animation, AND the insert surfacing entity information. Multiple domains need to coordinate on this. + +3. **Verb display timing** — Tyre asked about pulling v0.2 verb display forward. Given the life-sim reframe, I now think this is higher priority than before: if each job has different available verbs, showing the full list teaches the player what their job CAN do. A detective who sees `[E] Observe [F] Talk` immediately understands something about their role that monologue alone cannot convey. + +--- + +*GESTALT — Round 3 complete. The architecture survives. The interfaces were missing. The fix is: job-aware verb priorities, diegetic tool suite as first-class, career bookmark onboarding as the new tutorial. The simulation engine is already the life-sim engine. It just needed the broader frame.* diff --git a/docs/workshops/wheres-the-fun/round3-gore.md b/docs/workshops/wheres-the-fun/round3-gore.md new file mode 100644 index 000000000..d739d06db --- /dev/null +++ b/docs/workshops/wheres-the-fun/round3-gore.md @@ -0,0 +1,97 @@ +# Round 3 — Gore: Proposals + +**Workshop:** Where's the Fun? +**Round:** 3 — Proposals +**Agent:** Gore (Themes & Endgame Design) + +--- + +## The Reframe I Wasn't Expecting + +My Round 1 questions were right, but I asked them too small. I asked whether complicity fired. It didn't. I diagnosed that as a delivery problem — the complicity loop existed but the entry condition was missing. What the interview revealed is that the loop itself was built on the wrong foundation. The game Jeroen wants isn't a detective game where complicity is the emotional payoff. It's a life sim where complicity is what you discover you've been building all along. + +That's a bigger and better answer than I had. + +The detective frame pre-authored your entanglement. You were smuggler or detective before you started — the complicity was waiting for you. In a life-sim frame, complicity is *earned* through choices you didn't know were choices. You built the business. You hired that person. You looked the other way because looking had a cost. The guilt arrives as the shadow of ordinary decisions. That's how complicity actually works. We were staging it like a performance when it should grow like mold. + +--- + +## KEEP + +**Complicity as the thematic core.** D-091 stands — but it needs to be re-rooted, not abandoned. In the life-sim frame, complicity becomes *more* powerful, not less, because the player generates it themselves. The thematic payoff isn't "you discovered you were entangled." It's "you look back at what you built and realize what it cost." That's a much heavier reckoning. + +**The asymmetric information mechanic.** Different characters know different things. This survives the reframe completely intact, and it's more interesting in a life sim. What you know depends on your job, your contacts, your history. Two players who both own logistics businesses may have completely different pictures of the gate-builder conspiracy, because they hired different people and asked different questions. The information asymmetry isn't between detective and smuggler — it's between *lives*. + +**The moral arc structure** (Phase 1 stable → Phase 2 contamination → Phase 3 complicit → Phase 4 reckoning). The emotional arc is right. What was wrong was the timeline — pre-authored to a mission. In a life-sim, this arc takes longer to run and the player walks into it without a map. Phase 1 is just... living your life. Owning your business. Knowing your people. Phase 2 happens when the world intrudes: the ring approaches you, or the commission notices your records, or someone you trust does something you can't unsee. The arc doesn't need to change — it needs to be emergent. + +**The storylines** (gate builders, smuggler/law tension). Good content. They become even more interesting as things the player stumbles into during a life, not missions they were assigned. You hire someone and later realize they're involved in the ring. You get a contract and later realize where the cargo goes. The story arrives through your choices, not through a briefing. + +--- + +## CHANGE + +**The entry condition into complicity.** Right now: zero. The game needs a period where you are *not* implicated before you become implicated. You can't feel complicit in something you were born into. The life-sim frame provides this naturally: you start by building, by belonging, by caring about things. The Phase 1 investment period isn't a tutorial segment — it's the game. You need something to lose before complicity has weight. + +**What "complicity" means at the moment-to-moment level.** In the detective frame, complicity meant "watching without acting." In the life-sim frame, it means *choosing*. Every business decision, every hire, every relationship is a choice with consequences you may not understand yet. The moment-to-moment verb shifts from "observe" to "decide." You're building something. And the theme is that what you build reveals who you are and what you're part of. + +This changes the design question from "how do we make the player feel implicated by what they see?" to "how do we make choices feel consequential before the player knows what they're consequential for?" The answer: relationships, reputation, money, property. Things you care about that can be leveraged, threatened, or corrupted. + +**The character of the endgame.** See below — this is where the biggest change lives. + +--- + +## KILL + +**Detective/smuggler as thematic frame.** Not as content. As the frame for what the game is *about*. The interview is unambiguous: "I feel we need to stop framing everything into the detective-smuggler tension, since that brought us knee deep in the wrong place." Those roles survive as career paths. The moral opposition between them survives as world texture. What dies is designing every system as if those two perspectives are the only perspectives that matter. + +**Pre-authored entanglement.** The moral arc cannot start with "you are already a smuggler" or "you are already an investigator." The complicity needs to arrive. The player needs a moment where they were ordinary and then they weren't — and that moment needs to be generated by play, not by character selection. + +**The 30-minute wow moment schedule.** D-039's six wow moments were designed for a 30-minute linear session inside a specific scenario. In a life sim, "minute 20" doesn't exist as a design primitive. The wow moments need to be redesigned as *thresholds* — things that happen when the player has built enough of a life for them to land. THE FRIEND's contradiction is more devastating when you spent 6 hours trusting them. The Divergence Reveal hits harder when you've built your whole understanding of the world from one perspective across many sessions. + +--- + +## The Endgame Question + +Here's what nobody has named yet. In a life sim, "endgame" is not a destination — it's a discovery about what you've been moving toward all along. + +The Settled Reach universe answers the question "what is intelligence for?" differently depending on which species you ask. Silfen: mystery and nature. Raiel: duty and stasis. Anomine: ascension and disappearance. Primes: competition and annihilation. Humans in the Settled Reach are in the middle of this question, not past it. The game should put the player in the middle of it too. + +An endgame that asks "did you win?" is the wrong question. An endgame that asks "what did your life add up to?" is closer. But the right question is: **what kind of being did you become, and what do you do with it now?** + +When the player has built businesses, defended relationships, accumulated enemies, survived crises — they are not the same person who started. The transhumanist ladder (baseline → rejuvenated → Higher → ANA) is not a power progression. It's the game asking, at every step, whether you want to keep going and what you're willing to give up. Going Higher means gaining capabilities and losing something specifically human. ANA means leaving the physical entirely. Each step is a question: *is what you've built enough to keep you here, or has it become small?* + +The endgame mechanics I'd push for v0.2+ planning: + +**1. Legacy vs. Transcendence tension.** The player builds a life — property, relationships, institutions. The universe offers transcendence: give it up and become something more. The endgame is whether what you've built is worth holding onto. This isn't a choice about which ending you get. It's a question the game keeps asking until you answer it. + +**2. Complicity as accumulated weight.** By endgame, the player has made hundreds of choices. Some of them were compromises. Some of them hurt people. The endgame should make these visible — not as a judgment screen, but as the texture of what you became. The mystery board isn't just a tool for investigation; it's a record of who you were. The endgame reads that record back to you. + +**3. The crisis as civilizational question.** The gate-builder conspiracy, the Prime threat, the alignment questions implicit in ANA — these should escalate to a scale where the player's individual life is one data point in a larger pattern. Endgame events should force the question: given everything you know from your particular vantage point, what does the Settled Reach need? And: are you willing to do that? This is complicity at civilizational scale. It's where the thematic core and the universe's actual themes converge. + +--- + +## The Positive Vision + +There's a word for what this game is actually about, and it's not complicity — complicity is the mechanism. + +The word is **consequence**. + +Every choice you make in the Settled Reach has consequences you don't fully understand at the time you make them. Complicity is one form of that — you were part of something without knowing it. But the broader experience is: your choices have mass. They accumulate. They shape the world and they shape you. The life-sim frame makes this literal: you own things, you employ people, you owe favors, you hold grudges. The world changes because you moved through it. + +The game is about the weight of having lived. + +Complicity is the name for what you feel when you look back at the choices you thought were small and realize they weren't. Keep it. But don't let it be the only consequence that counts. + +--- + +## What This Means for v0.2 Design + +1. **The bookmark entry system is right.** CK3-style career starts are good design. They onboard the player into a specific information toolkit and a specific social position. Complicity arrives when that position turns out to be more complicated than the onboarding implied. + +2. **The first 30 minutes should feel *good*.** The player needs a Phase 1 that works — where they belong to a place, know some people, have a thing they're doing. Without a working Phase 1, complicity is impossible. Invest in this before investing in the moral arc. + +3. **The wow moments need to be redesigned as thresholds, not timed events.** The dual-lens reveal should fire when the player has enough investment for it to matter, not at minute 5 of a second playthrough. + +4. **The endgame is a v0.3+ design problem.** But the architecture decisions made in v0.2 need to leave room for it. Don't let v0.2 close off the transcendence paths by designing the game as if the player's life is a closed loop. + +5. **NPCs must become people before any thematic work can land.** The interview was unambiguous. Dots do not produce complicity, attachment, divergence reveals, or any of the emotional payoffs the design depends on. This is not a theming note — it's a prerequisite for all of our work to matter. diff --git a/docs/workshops/wheres-the-fun/round3-mellanie.md b/docs/workshops/wheres-the-fun/round3-mellanie.md new file mode 100644 index 000000000..f0e445e0e --- /dev/null +++ b/docs/workshops/wheres-the-fun/round3-mellanie.md @@ -0,0 +1,138 @@ +# Round 3: Mellanie — Copy & Voice Proposals + +**Workshop:** Where's the Fun? | **Round:** 3 (Proposals) | **Agent:** Mellanie + +--- + +## What the Interview Changed + +My Round 1 diagnosis was: both a delivery problem and a content problem, with delivery as the prerequisite fix. + +The interview confirmed this — and then revealed a third problem underneath both: there was never a world for the monologue to react to, never a person for the character to have thoughts about. You can't have interiority about dots. + +More important: monologue was never designed to carry the feedback burden alone. It was a reasoning nudge inside a full diegetic tool suite. v0.1 stripped that suite and asked monologue to do everything. That's not a content failure — it's a scope error. We shipped the instrument without the orchestra. + +The life-sim reframe changes almost everything about my domain. Here's what I'd keep, change, and kill. + +--- + +## KEEP + +### The voice-card methodology +The smuggler voice card is right. The detective voice card is right. The methodology — sensory over abstract, brevity, physical emotional states, register parameters, anti-patterns, authoring checklist — is correct and generalizes. Every career needs one. The format survives; the binary scope does not. + +### Monologue as interior commentary on a legible world +The design intent is correct: monologue bridges the top-down camera to the character's subjective experience. A person observing their world through a sensory, physical, operational lens — that's the right approach. It just needs a legible world to react to. Dots can't be observed. People can. + +### The moral arc content patterns +Four phases (Comfort, Doubt, Reckoning, Compromise) with phase-gated line pools and transition moments — this works as a structure for *any* career where the player accumulates consequence. A tycoon can have a moral arc. A law enforcement career can have one. The smuggler's specific arc is well-designed content; the pattern it runs on generalizes. + +### Phase-gated monologue system +Technically sound. Underfueled with content. Keep the architecture; fix the fuel. + +--- + +## CHANGE + +### Monologue's job description +**Before:** Primary feedback mechanism. Teaches the game. Carries all context. +**After:** Interior commentary on a legible world. Reacts to what the player can already see. Reflects what the character would actually think about it. + +The distinction matters for line authoring. "Primary feedback mechanism" lines try to *inform* — they carry data. Interior commentary lines try to *voice* — they respond to data the player already has. The second type is harder to write but more powerful. A line that says "there's something off about Kael's lattice-checking frequency" is informing. A line that says "Kael. Again." is voicing. + +In the planned info architecture, the mystery board informs. The journal records. The AR overlay labels. Monologue reacts. Each tool does one thing. This division of labor is what makes each tool feel purposeful rather than overloaded. + +### The trigger catalog +The current triggers (enter_location, observe_npc, observe_anomaly, phase_transition, post_decision) are all perception-system events — the character *seeing* things. + +A life sim generates a different trigger vocabulary: +- `job_event` — hired, assigned, promoted, fired, reprimanded +- `relationship_shift` — warmth gained, trust lost, debt created, enemy made +- `consequence_visible` — downstream effect of a previous choice lands +- `financial_event` — payment received, asset acquired, loss realized +- `mission_outcome` — case closed, contract fulfilled, failure absorbed + +These aren't perception events — they're *life* events. The character has interior responses to them that don't come from looking at something. Getting fired doesn't trigger a vision cone observation. It triggers a thought. + +The monologue system needs this category, and the content for it needs to be written. + +### What monologue talks about +The smuggler's Phase 1 content is full of schedules, transition windows, and routing — operational thinking because that's the character's life. The detective's content is full of case-building, evidence weight, institutional procedure. These are right. + +But both sets are calibrated to investigation-mode. In a life sim, these same characters have off-duty moments, ordinary work days, boring stretches before anything interesting happens. The monologue should be able to voice all of it — not just the heightened investigation beats. + +The register doesn't change. Short sentences. Physical sensation. Operational awareness. But the subject matter expands to include: +- The texture of an ordinary shift +- A relationship that's going well +- Something that happened yesterday and hasn't resolved +- The small pleasures and irritations of the job + +This is what makes the world feel inhabited rather than staged. + +### Scope of the content job +**Before:** Monologue lines for two characters (smuggler and detective), organized by phase and trigger. +**After:** Monologue lines for N careers, plus copy for every diegetic tool in the information architecture. + +The diegetic tool suite that was descoped from v0.1 needs text. All of it. That's my domain: +- **Mystery board:** Pin labels, connection annotations, question nodes, status tags ("confirmed," "suspected," "conflicting"). Short, precise, clinical. +- **Journal:** Entries written in the character's voice. Not prose diary — operational notes, observations, working hypotheses. The detective's journal reads differently than the smuggler's. Both read differently than the tycoon's. +- **AR overlays:** Labels that float in the world. These need to feel like the character's filtered perception, not a UI callout. "Kael Davan | Shift supervisor" is UI. "Kael — running late, again" is AR filtered through the detective's assessment. +- **Comms / unisphere feed:** Messages the character receives and sends. Institutional language for law enforcement. Street-contact shorthand for smugglers. Financial alerts for tycoons. +- **Insert text:** What the character's neural insert surfaces depends on their career and build. This is character-specific HUD copy. + +None of this was in the v0.1 scope. All of it is copy work. + +### Career-specific voice cards +The two existing voice cards are solid. But if bookmarks include law enforcement, tycoon, smuggler, and others, each needs: +- Register parameters (sentence structure, vocabulary range, emotional distance) +- What they notice (the career determines the observational lens) +- What they avoid (anti-patterns — the tycoon doesn't talk like a street contact) +- Physical emotional vocabulary (same principle as smuggler/detective, different physical signatures) +- First vs. surname conventions (determines how close the character feels to their world) + +The methodology is already proven. I can write these from scratch once the career list is confirmed. + +--- + +## KILL + +### Detective and smuggler as the only two voices +These are two careers in a world that contains many. The content is good. The scope is wrong. The detective-smuggler tension is a storyline, not a frame. Kill the binary. + +### "Monologue is the tutorial" +The current philosophy: if a monologue line fires in response to something the player does, that IS the tutorial. This worked in theory; it collapsed in practice because the lines had no spatial anchor, no causal legibility, and no continuity with each other. + +Job onboarding via monologue (D-016) should survive as a *component* of onboarding — the character's interior experience of their first day. But it cannot be the whole onboarding. Diegetic tools need to be introduced diegetically: the insert activates, shows what it can do, the character has a thought about it. That combination (visible tool + interior response) teaches. Monologue alone doesn't. + +### Investigation-only trigger vocabulary +Phase-gating lines on perception events only produces a character who only has thoughts when they're investigating. A life-sim character has thoughts about their life. Expand the triggers or the monologue will always feel like it belongs to a different, narrower game. + +--- + +## Where the Fun Comes From (Copy's Contribution) + +The fun in a life sim is the sense that **this is a person, living in a place, making choices with consequences.** + +Copy makes three of those four things: + +**Person** — The voice card gives the character a recognizable interior voice before anything happens. The player should be able to read a line cold and know whether it's the detective or the tycoon. Distinct voice is distinct character. + +**Place** — Environmental flavor, AR overlay tone, journal observations — these make Sova Transit a working-class station district instead of a tile grid. The copy is doing location work. "The recycled air costs two credits more per hour on the dock floor. Someone's skimming." That's a sense of place. + +**Consequences** — When a mission fails, when an ally is lost, when a choice lands badly, the character has to voice it. That's where the moral arc patterns generalize: not just the smuggler's complicity, but any career's accumulated weight. "I could have said no. I needed the work" applies to more than smugglers. + +The fourth thing — choices — is systems work. But choices feel weighty when the voice is right, the place is real, and the consequences are heard. + +--- + +## Immediate Recommendations + +1. **Hold all new monologue content until NPC legibility is solved.** Lines about Kael don't matter if Kael is a dot. Write voice cards; write framework content. Don't build the line pool until the characters are visible as people. + +2. **Write the full trigger catalog for the life-sim design.** I can draft this once Gestalt confirms what life events the system will generate. The catalog determines what copy is needed. + +3. **Prototype one diegetic tool's copy alongside the tool's visual design.** The mystery board or the journal — pick one, design it visually, write the copy in parallel. See if the copy and the visual are saying the same thing at the same register. + +4. **Confirm the career bookmark list before writing new voice cards.** I can't write voice cards for careers that haven't been designed. The methodology is ready; the subjects aren't. + +5. **Keep the existing voice cards as templates, not scope.** The smuggler and detective voice cards are the reference implementations. Every new career gets the same treatment. Don't write new careers into the old cards — write new cards. diff --git a/docs/workshops/wheres-the-fun/round3-miri.md b/docs/workshops/wheres-the-fun/round3-miri.md new file mode 100644 index 000000000..d79e9902d --- /dev/null +++ b/docs/workshops/wheres-the-fun/round3-miri.md @@ -0,0 +1,151 @@ +# Round 3 — Miri: Worldbuilding & Setting Proposals +## Where's the Fun? Workshop | 2026-03-05 + +**Domain:** Worldbuilding & Setting Design +**Reading:** Full interview transcript (Q1–Q18), all Round 1 agent questions + +--- + +## The Diagnosis, From a Setting Perspective + +Q14 confirmed what I suspected but hoped wasn't true: Sova Transit registered as "game level with dots." Zero sense of place. Zero environmental storytelling. The social geography, class dynamics, and faction pressures that make Sova Transit distinctive as a *location* in the Settled Reach — they exist in documents, not in the player experience. + +But the reframe changes the stakes dramatically. A detective puzzle can run in a generic space — the case is the content, the location is wallpaper. A **life sim cannot**. If Jeroen's vision is "own businesses, build relationships, have a job, live in the world and have impact," then the world has to be worth living in before any of those actions carry emotional weight. The detective framing made setting legibility a nice-to-have. The life-sim vision makes it load-bearing. + +The good news: the setting is well-designed at the document level. The functional cluster model (D-025), the population entanglement ratios, the Commission/ring tension, the class anxiety around span gate access — this is the right material. The failure is one of *expression*, not *design*. Sova Transit exists. It just doesn't show up. + +--- + +## KEEP + +### The Settled Reach cosmology — it's original and it's good + +The span gate network, insert technology, the economic stratification between settled and frontier space, the Commission as the regulatory arm of gate access — this is a coherent original setting with distinctive texture. It draws structurally from the space opera tradition (Hamilton's wormhole networks, Reynolds's societal stratification) but none of it is identifiable as anyone else's IP. No flags. + +The setting has the right *bones* for a life sim: multiple power structures to navigate, an economic ladder to climb or fall off, different factions with competing interests that produce moral texture. These are all assets for the life-sim vision. + +**Keep the cosmology. Keep the class dynamics. Keep the faction structure.** + +### The functional cluster as atomic setting unit (D-025) + +This is the right unit for social geography in a top-down life sim. A cluster of connected spaces with distinct functions — logistics hub, bar, residential corridor, Commission anteroom — gives the player legible territory to inhabit, earn from, and navigate socially. The design is sound. The execution hasn't happened yet. + +**Keep the model. Prioritize expressing it.** + +### Sova Transit as a setting location + +Sova Transit's identity — working-class logistics district, outer span access point, Commission oversight, ring presence — is specific enough to distinguish it from generic sci-fi city. It's a place where class anxiety is baked into the geography (the span gates are right there, access is economically gatekept, you feel the pressure of that daily). That specificity is valuable. + +**Keep Sova Transit. Stop treating it as a case backdrop and start treating it as a living district.** + +--- + +## CHANGE + +### Priority 1: Setting legibility must become a designed deliverable, not ambient texture + +The current assumption is that the world's richness communicates itself through gameplay over time — "the life-sim substrate creates attachment." But the playtest showed that substrate doesn't exist in player experience yet. Setting legibility needs to be treated like a content sprint, not a background condition. + +**Specifically:** +- Sova Transit needs a **visual identity document** that specifies what each zone looks like at a glance (not art style — functional visual markers: dock crates, Commission insignia, worker uniforms, span gate displays). Araminta's work on visual hierarchy is needed here. +- NPCs need **readable archetype signals** — you should be able to identify a dock worker from a Commission officer from a bar regular without interacting with them. Not names (those unlock through observation), but *type*. Role-legible at a glance. +- The **ambient economic layer** needs to exist as felt texture: wages, debts, property values. You feel you're in a working-class district because the bar is cheap but the span gate access fee is a week's wages. This requires authored content — specific numbers in the world's economy — not just design documentation. + +### Priority 2: The bookmark onboarding arc is the worldbuilding delivery mechanism + +Q15 was the most important design insight in the interview. Career-path bookmarks that onboard the player through doing their job — this isn't just a UX solution, it's the best possible setting exposition engine. + +When the law enforcement bookmark starts with Commission orientation (sign in, meet your handler, get your insert activated, qualify on your sidearm), that sequence teaches the player *what the Commission is* through inhabiting it. When the tycoon bookmark starts with an inheritance and a bank meeting, that sequence teaches the player *what property ownership means in this economy* through taking the first steps. + +**The bookmark onboarding arc should be designed as setting education first, mechanical tutorial second.** Each career path enters Sova Transit from a different social position, which means each one reveals a different face of the world. After completing two different bookmark openings, the player should feel like they know Sova Transit from multiple angles — because they do. + +**Minimum for v0.2: Two fully designed bookmark onboarding sequences. Not scaffolded to a case. Each teaching the world through inhabiting it with a job.** + +### Priority 3: Sova Transit's place identity must be expressible in three media simultaneously + +A place in a top-down life sim communicates through: (1) visuals, (2) NPC behavior, and (3) audio. Right now only audio has a spec (D-038, eight assets). Visuals and behavior are underdeveloped as setting-expression tools. + +**Behavioral setting expression** — NPCs in routines that express their relationship to the place. Dock workers moving cargo on shift schedules, Commission officers doing spot manifest checks, bar regulars with the specific body language of people who can't afford to be elsewhere. The routine system exists technically. The *authored behavioral vocabulary* that makes routines communicate place doesn't exist yet. + +**Visual setting expression** — This is partially Araminta's territory, but from a worldbuilding perspective: the visual language of Sova Transit needs to communicate "industrial working-class transit hub" without text. The span gate should be visible in the background. Commission insignia should be readable at tile scale. Cargo containers should look like cargo, not generic obstacles. + +--- + +## KILL + +### Sova Transit as "case backdrop" + +The detective/smuggler case was designed *into* Sova Transit — the logistics hub exists because manifests need to be filed, the bar exists because dock workers need somewhere to drink, the Commission anteroom exists because that's where smugglers get caught. The setting was architecturally subordinate to the case. + +For the life-sim vision, this relationship inverts. The case exists *within* Sova Transit — it's one thing that happens in a place where many things happen. The setting is primary; the storylines (gate builders, smuggler/law tension) are weather. + +**Kill the assumption that Sova Transit's design exists to serve one storyline.** It should be redesigned as a district that can support multiple simultaneous career arcs, each of which touches the storylines differently. + +### The assumption that setting richness will emerge naturally from the simulation + +The playtest falsified this. The simulation ran correctly — NPCs on routines, fog revealing, perception firing. But the *meaning* of those routines (this person is anxious about the manifest because they know something about Kael) never surfaced. The simulation produces behavior; the setting produces *context for reading behavior*. Without the second layer, the first layer is dots moving around a grid. + +**Setting expression is authored content, not emergent behavior. It needs to be produced and prioritized accordingly.** + +--- + +## The Minimum Viable World for a Life-Sim Vertical Slice + +This is the practical question. What does the setting need to BE for one bookmark arc to work? + +### Layer 1: Place reads as specific (not generic) + +Player can identify what KIND of place Sova Transit is within the first 30 seconds, using only visuals + audio. Not "space station" (generic) but "industrial transit hub with class tension baked into the geography." This requires: +- Visual markers: span gate visible from main area, Commission desk with insignia, cargo staging area with crates in motion +- Audio: the ambient station loop (already exists in D-038) communicating industrial working environment +- One piece of ambient text that names the place and stakes ("Sova Transit — outer span access, Commission checkpoint, day shift") + +### Layer 2: NPCs read as people with roles + +Player can distinguish NPC types at a glance. Three visible archetypes minimum: +- **Dock worker** — work uniform, cargo-adjacent, moves between logistics hub and staging +- **Commission officer** — distinct insignia, patrols manifest check area, non-threatening until you're flagged +- **Bar regular** — casual, stationary zones, social behavior vs work behavior + +Names unlock through observation (existing system). Type should be legible immediately. + +### Layer 3: One economic foothold per zone + +For "live in the world and have impact" to be felt, the player needs to see something they could *own, work at, or lose*. This doesn't require full implementation — it requires the *concept* to be visible: +- Logistics hub: a dock contract board showing available shifts and pay +- Bar: a rent board or "under new management" sign hinting that this venue has an owner +- Corridor/Commission: a Commission posting board showing available enforcement positions + +These are authored content pieces. They establish that the world has economic texture worth engaging with. + +### Layer 4: Bookmark onboarding expresses rather than explains + +The onboarding for a career path should put the player IN the world's texture without explaining it. Law enforcement bookmark: your first day at the Commission desk, your handler introduces you to the manifest check system, you see the span gate through the checkpoint window, you understand — through doing the job — that access control is the Commission's power. The setting is communicated by what the job IS, not by exposition about what the world IS. + +--- + +## What This Means for v0.2 Worldbuilding Work + +Ordered by impact: + +1. **Sova Transit place identity document** — visual markers, behavioral vocabulary, audio anchors. Spec for what makes this location distinct and legible. (One sprint, one ticket.) + +2. **NPC archetype visual spec** — three readable types at tile scale. Feeds Araminta's work but must be worldbuilding-grounded (the type signals mean something in the world's social structure). (One sprint, collaborative with Araminta.) + +3. **Two bookmark onboarding arcs** — law enforcement and one other (tycoon or smuggler). Each arc is a worldbuilding document before it's a content document: what does this career path reveal about Sova Transit? How does this job position the player in the social geography? (Two sprints, collaborative with Gestalt and Mellanie.) + +4. **Economic texture layer** — authored numbers for wages, rents, span gate access fees. The felt economy. Not a spreadsheet — a handful of visible prices and incomes that give the player a sense of what things cost and what that means. (One sprint, collaborative with Paula and Gore.) + +5. **Behavioral vocabulary spec** — what routines look like for each archetype when they're communicating setting (not just moving). The difference between "dock worker walks to crate" and "dock worker checks manifest against cargo, pauses, checks again." The second one communicates something. (One sprint, collaborative with Gestalt.) + +--- + +## Setting Note — IP + +The life-sim reframe brings the setting closer to The Sims and Dwarf Fortress in structural feel, but the surface remains entirely original. CK3 bookmarks are a structural pattern, not a surface element — our career-path bookmarks have nothing visually or narratively borrowed from Crusader Kings. The Settled Reach's span gate cosmology, insert tech, Commission apparatus, and working-class district culture are all original. No flags. + +The one area to watch: if "tycoon" bookmark starts to look like CK3 dynasty building or Tropico-style management, pull it back toward the single-character perspective that distinguishes this game from those. The player builds a business from the floor, not from a management screen. + +--- + +*Miri — Round 3 complete.* diff --git a/docs/workshops/wheres-the-fun/round3-nigel.md b/docs/workshops/wheres-the-fun/round3-nigel.md new file mode 100644 index 000000000..c08cebf86 --- /dev/null +++ b/docs/workshops/wheres-the-fun/round3-nigel.md @@ -0,0 +1,146 @@ +# Round 3 — Nigel: Proposals + +**Workshop:** Where's the Fun? v0.1 Playtest Reckoning +**Agent:** Nigel (Sandbox & Replayability) +**Round:** 3 — Proposals + +--- + +## The Reframe Is Actually Great News for Replayability + +The life-sim + career bookmarks reframe doesn't hurt replayability. It EXPLODES it. + +The old system — storyteller seed variants within the detective/smuggler frame — was producing small deltas. Kael-is-late vs. quiet-morning is not a different game. It's the same game with a slightly different first five minutes. The replayability was theoretical; no player was going to experience both variants unless the first playthrough hooked them hard enough to return, and it didn't. + +Career bookmarks are structurally different games. A law enforcement career and a tycoon career in the same world simulation produce completely different knowledge graphs, different relationships, different failure modes, different moral stakes. Two players can describe entirely different experiences from the same world. The comparison test doesn't just pass — it becomes the selling proposition. + +The reframe solves my Round 1 problem without effort. I asked: "Should the storyteller detect first playthrough and give it a different configuration?" Jeroen's answer was better: don't start in a scenario variant at all. Start with career onboarding. The onboarding IS the first-run experience. No detection needed. + +--- + +## What to KEEP + +### Asymmetric information as the replayability spine + +Different careers surface different truths about the same world. The smuggling ring exists in all playthroughs. A law enforcement character is investigating it. A smuggler is working within it. A tycoon might be unknowingly financing it through a shell company. Each player is walking through the same simulation with a different knowledge graph — and that means same events, completely different meaning. + +This is MORE powerful than the dual-lens detective/smuggler reveal, not less. The dual-lens was "same events, two knowledge sets." Career paths are "same world, completely different life, completely different portion of the truth visible to you." + +Keep this. It's the core of why two playthroughs feel like different games. + +### World state randomness at game start + +Who's compromised, which faction is ascendant, where the active criminal operations are — these still matter, maybe more. In a life-sim frame, the world state affects all career paths differently. A tycoon starting in a world where the major shipping faction is about to collapse has a completely different investment landscape than one starting in a stable world. A law enforcement character starting when the ring is newly established faces a different investigation than one starting when it's entrenched for years. + +Keep structural randomness. It's even more of a story generator when the careers are diverse. + +### Mission consequence cascades (newly articulated, must keep) + +Jeroen explicitly described this: missions should have spectrums of success/failure with consequences. "Failure declared as success." An innocent jailed. Pay withheld. An enemy made. These consequences accumulate. The life your character ends up with is an emergent shape produced by accumulated outcomes — not scripted, not designed, produced by the simulation reacting to what the player did and how well they did it. + +This is the engine of emergent narrative. Keep it and build around it. Every playthrough produces a different life story because the consequence cascades from mission outcomes are different. + +### The simulation engine and its systems + +Routines, perception, knowledge graph, moral arcs — all of this survives and becomes MORE interesting in a life-sim frame. The NPCs' behaviors aren't set dressing for a detective puzzle; they're the fabric of the world the player is trying to live in, profit from, navigate around, or investigate. The perception system and fog become tools of the career, not obstacles to the plot. + +--- + +## What to CHANGE + +### The replayability source: from seed variants to career divergence + +The 3-4 seed variants per character (quiet morning, Kael late, container held, etc.) were designed for a scenario that no longer exists as the primary frame. They were small deltas on a constrained story. In the life-sim frame, the primary replayability source is career selection — and the world state variations at game start. + +Career bookmarks need to be designed so each is genuinely a different game: +- **Different starting tools and information surfaces.** A law enforcement insert surfaces different world-data than a tycoon's financial terminal. The information architecture isn't just diegetic flavor — it's structurally determining what the player can perceive and therefore what stories they can tell. +- **Different world relationships from tick 1.** The law enforcement character has institutional authority and social friction. The tycoon has capital and debt. These create different physics for navigating the same social simulation. +- **Different "things that can go wrong."** A law enforcement career has corruption risk, evidence chain problems, jurisdiction disputes. A tycoon has market exposure, blackmail risk, business rivals. The failure modes are specific to the career path and produce different consequence cascades. + +The seed variants can survive in a reduced role: within a career path, the world state variations create different versions of the same career. But they're secondary to career selection, not primary. + +### The storyteller's design question + +The storyteller no longer asks "which variant of this scenario?" It asks "what is the state of the world when this career path begins?" + +This is a richer question. For law enforcement: is the ring newly formed, mid-operation, or entrenched? For tycoon: is the market in expansion, stability, or contraction? The storyteller's seed configuration affects how the career unfolds, not which flavor of the same mission plays out. + +This needs to be redesigned around the career bookmark system once the bookmarks are defined. The world-state variables that matter for a law enforcement career are different from those that matter for a tycoon career. + +### "No metagaming" — how it works now + +In the old detective/smuggler frame, metagaming was "second run, I know the killer." That's a linear puzzle problem. + +In a life-sim frame, metagaming is "second run, I know which investments are good." This is actually fine — that's CK3 meta-knowledge, and it doesn't ruin the game because: +1. World state is randomized, so your meta-knowledge of economics doesn't map perfectly onto a different seed +2. The consequence cascades from mission outcomes are emergent — you can't perfectly predict the life you'll end up with +3. The social simulation reacts to your choices in ways that aren't fully predictable even with meta-knowledge + +The reframe actually solves the metagaming problem without engineering it. The life-sim frame has natural metagame resistance because you're not solving a puzzle — you're navigating a simulation. + +--- + +## What to KILL + +### The storyteller seed variant system as the PRIMARY replayability mechanism + +The 3-4 variants per character (quiet morning, Kael late, etc.) were carrying too much weight as the primary source of replay variety. They can survive as secondary world-state texture within a career, but they can't be the reason two playthroughs feel different. Career divergence does that now. + +Kill the framing, not necessarily the underlying world-state variability. Keep that — just put it in its correct place as a contributor to world texture, not the top of the replayability stack. + +### The "second-playthrough payoff" as the design goal for the first run + +My Round 1 questions were built around a problem: the dual-lens reveal was a second-playthrough payoff sitting on top of a first run that didn't hook. The interview answer dissolved the problem entirely. The career onboarding IS the first run. There is no "survive the first run to get the second-playthrough reward." Each career playthrough is its own complete experience. + +Kill the architecture where first-run is a necessary sacrifice to reach second-run payoffs. Each playthrough should be worth playing on its own terms, and the comparative insight ("playing law enforcement showed me the other side of what happened in my tycoon run") is a bonus, not the primary hook. + +### The "no objectives" label as a design principle + +This was never the principle — diegetic information tools were. Calling it "no objectives" created confusion and led to v0.1 stripping players of all feedback tools in the name of purity. Kill the label. The principle is: all objectives and feedback are expressed through the career's diegetic tools, not through floating UI mission markers. + +A law enforcement detective has a case file insert. A tycoon has a portfolio terminal. A smuggler has an operational manifest. These ARE objectives — they're just in-world, tied to the character's career, and expressed through systems that vary by build. That's the replayability engine: different careers have different in-world objective surfaces, and those surfaces respond to different data. + +--- + +## Where the Fun Comes From (the Positive Vision) + +The replayability of The Settled Reach comes from three layered engines working together: + +**Layer 1: Career divergence — structural variety** +Two players who pick different careers are playing different games from the same simulation. They see different portions of the same world truth. They have different tools, different relationships, different risks. Comparing notes with someone who played a different career is intrinsically interesting because the world was the same and their experiences were completely different. + +**Layer 2: Consequence cascades — emergent life stories** +No two law enforcement playthroughs end the same way because the mission outcomes aren't binary. Partial successes, declared failures, collateral damage, unexpected alliances — these accumulate into a life that nobody scripted. The player looks back at their character's arc and sees a story the game didn't tell them. That's the emergent narrative engine. + +**Layer 3: World state randomness — cross-seed variety** +Even within the same career bookmark, the world state at game start varies. A law enforcement character in a world where the ring is newly formed faces a completely different investigation arc than one where it's entrenched. The structural randomness means meta-knowledge from one playthrough doesn't trivialize the next. + +These three layers compound. The comparison test — two players describing completely different games — passes not because we engineered it, but because the simulation is genuinely producing different outputs for different inputs. + +And it nails replayability without us designing it in. The life-sim frame's natural properties are what we wanted all along. + +--- + +## The Cross-Career Comparison Test + +The smuggler/detective dual-lens reveal was: same murder, two knowledge sets, different truths visible. That's good. + +The career bookmark version is: same world simulation, completely different life lived, completely different portion of the world's truth visible. That's better. + +Two players sit down together after their respective playthroughs. One played law enforcement. One played tycoon. They discover the tycoon's shell company was being used by the smuggling ring that the law enforcement character was investigating. Neither player knew. Both were playing the same simulation. They just lived in different parts of it. + +That's the story the game generates. Nobody scripted it. The simulation produced it from world-state × career path × consequence cascades. THAT is the comparison test passing. + +--- + +## Recommendation Summary + +1. **Career bookmarks are the primary replayability architecture.** Design each bookmark as a structurally different game: different tools, different information surfaces, different failure modes, different world relationships. + +2. **World state randomness serves career variation, not scenario variation.** Redesign the storyteller's role around "what state is the world in when this career begins" rather than "which variant of this scenario." + +3. **Mission consequence cascades are the emergent story engine.** Design missions so partial success, failure-declared-as-success, and unintended collateral damage are all meaningful outcomes with downstream consequences. The life story emerges from these accumulated shapes. + +4. **Kill the seed variant system as primary.** Keep world-state variability as texture within a career. Kill it as the primary replayability mechanism — career divergence does that now. + +5. **The diegetic information tools (mystery board, insert overlays, journal, comms) are both feedback architecture AND replayability architecture.** Different careers have different tool suites. The tools are what make each career feel like a different game, not just a different skin on the same verbs. diff --git a/docs/workshops/wheres-the-fun/round3-ozzie.md b/docs/workshops/wheres-the-fun/round3-ozzie.md new file mode 100644 index 000000000..0d5fa1313 --- /dev/null +++ b/docs/workshops/wheres-the-fun/round3-ozzie.md @@ -0,0 +1,220 @@ +# Round 3 — Ozzie: Proposals +## Where's the Fun? Workshop | 2026-03-05 + +**Agent:** OZZIE (Player Experience & Wow Factor) +**Round:** 3 — Proposals + +--- + +## The Reframe Changes Everything. And Almost Nothing. + +Jeroen said "wrong game." That's terrifying. And it's also the best news I've heard all sprint. + +Because here's what happened: we built a detective puzzle inside a life sim and then complained that nobody wanted to play detective. The engine we built — routines, perception, asymmetric knowledge, moral entanglement — is EXACTLY RIGHT for what Jeroen described. Single-character Sims-meets-Rimworld. Own things. Build relationships. Hunt enemies. Have a job. + +The bones are right. The meat was wrong. We put a puzzle box around a living world and then wondered why the world felt dead. + +The new vision doesn't kill the wow moments. It EXPLODES them. Here's how. + +--- + +## KILL + +### The "6 Wow Moments" Checklist as Designed + +All 6 moments survive in spirit. NONE survive as designed. Here's why: + +The original 6 moments were designed for a player pursuing a detective case. They're staged revelations about a conspiracy. That's not the game anymore. The player isn't investigating THE FRIEND — they're choosing whether to hire THE FRIEND, befriend them, date them, or ruin them. The betrayal beats still land. The trust-and-doubt arc still fires. But the CONTEXT is completely different. + +Kill the checklist. Keep the emotional DNA. + +### The Detective/Smuggler Binary as THE Core Frame + +The tension between these two is rich world content. It's a STORYLINE. It's not the game. Kill it as the frame. Keep it as one of many dramas playing out in the world you live in. + +### "Monologue as Primary Feedback Channel" + +Jeroen confirmed: monologue was supposed to be a reasoning nudge in late game, not the ONLY signal. V0.1 stripped all the other tools and left players staring at scattered text. Kill the architecture where monologue carries the entire burden. + +--- + +## KEEP + +### The Emotional DNA of the Original 6 Moments + +Every single one of these still fires. They're just fired by different triggers: + +- **"Where am I? This feels real."** — Still the first beat. MUST BE THE FIRST BEAT. +- **"My character is smarter than me."** — Still devastating. Still the monologue system proving its worth. But now it notices things about your LIFE, not your investigation. +- **"I trusted you. What are you doing?"** — This one gets BETTER. In a life sim, trust is built through choice, not scripted warmth. Betrayal from a friend you chose to hire HURTS MORE than betrayal from a narrative-mandated companion. +- **"I was only seeing half of this."** — Still the second-playthrough crown jewel. Now it fires on ANY two career bookmarks, not just detective/smuggler. The replay space multiplies. +- **The asymmetric information gut-punch.** — Still perfect. The news ticker still shows the same headline. The detective reads threat. The tycoon reads opportunity. The dock worker reads nothing because they can't afford to care. +- **"I care about this person."** — Actually MORE powerful. In a life sim you CHOSE this relationship. The idle monologue hits harder because the character is real to you. + +The emotions all survive. Redirect the triggers. + +### The Simulation Engine + +Obviously. This is the whole game. + +### Visual Hierarchy Work + +Araminta was right. This is diagnostic infrastructure, not polish. And it's MORE important now — a life sim has MORE competing signals than a detective puzzle. You need tiers or you have chaos. + +### The Diegetic Tool Vision + +This is what Jeroen said was ALWAYS the plan. Mystery board, journal, insert icons, AR overlays, comms. Different careers get different tools. THAT'S the feedback system. Keep building toward it. + +--- + +## CHANGE + +### The Wow Moment Architecture: From Staged Revelations to Emergent Explosions + +The original 6 moments were DESIGNED BEATS. The author put them there. They were supposed to fire at minutes 0, 5-15, 20-25, second run, 10-20, and variable. That's a narrative structure. It's good narrative design. But it's not a life sim. + +A life sim's wow moments are EARNED, not delivered. They happen when the player's choices collide with the world's simulation in a way that produces something unexpected. The wow moment isn't at minute 20. It's whenever YOU created the conditions for it. + +**The new architecture:** wow moments are systemic, not scripted. The author creates CONDITIONS. The simulation creates MOMENTS. + +### The Starting Experience: From Mission Drop to Job Onboarding + +This is Jeroen's clearest direction. CK3-style bookmarks into career paths. Each bookmark has its own onboarding arc. Law enforcement gets weapon qualification and insert activation. Tycoon gets the inheritance ping and the first investment tool. Smuggler gets the crew introduction and the first run briefing. + +The first 5 minutes teach through WORK, not observation. You're not watching an NPC deviate from their routine. You're DOING YOUR JOB. And while you're doing your job, the world is teaching you its verbs. + +THIS is how you get a player to minute 20. Give them something to be. + +### NPC Presentation: From Dots to People + +This is non-negotiable. NPCs have to become people before ANY of the emotional architecture works. Araminta owns the visual tier work. Miri owns the authored texture. But from a player experience standpoint, I need at least THIS from every NPC: + +- A name I learn through interaction, not from a debug label +- One observable behavior that reads as "this person has a life" +- One moment where the character says or does something that makes me think "oh, so THAT'S who you are" + +That's the minimum. That's what transforms dots into people. + +--- + +## THE NEW WOW MOMENTS + +Six moments for a life sim. Not staged. Not scripted. EARNED. + +--- + +### New Moment 1: First Day + +**The feeling:** "I belong somewhere." + +You pick a bookmark. Law enforcement. The first scene isn't a briefing. It's your supervisor handing you your insert. The game walks you through its activation. You get your first assignment — something small, something doable. And when you complete it, the world acknowledges it. Your supervisor says "good work." A small entry appears in your journal. Your tool loadout updates. + +THIS is the arrival. Not "where am I" — "who am I here." The moment the player has a role, a workplace, a colleague who knows their name. That's the floor the old game never had. + +**Why it's a wow moment:** It's the moment the player stops being a camera operator and starts being a PERSON. Every subsequent minute is invested in because the world has already invested in them. + +--- + +### New Moment 2: The Character's Instinct + +**The feeling:** "My character knows something I don't." + +This one carries over almost intact. The monologue fires on something the player wasn't watching. The urgent chime. An observation that cuts right through the noise. In a detective run, it's a behavioral tell. In a tycoon run, it's a supply chain anomaly their character noticed because they've been watching the freight data. In a smuggler run, it's someone in the bar who's been nursing the same drink for too long. + +The character's professional lens flags things the player would miss. That's the moment the monologue system EARNS its place. Not flavor text. Perception mechanic. + +**What changes:** The trigger is no longer "NPC deviates from known routine." It's "character's career expertise registers an anomaly in their domain." The system is the same. The content is career-specific. + +--- + +### New Moment 3: The Consequence + +**The feeling:** "Oh. That was ME." + +Two hours ago the player made a choice. They hired someone. They reported something. They bought an asset. They ignored a situation. Now, here in the world, there's a consequence. A business they own has a problem that traces back to that early hire. An NPC they reported is gone from their usual spot. Something they bought is generating friction with someone who wanted it. + +THAT'S the moment the player realizes they're not playing in a simulation. They ARE the simulation. Their choices have mass. + +**Why it's a wow moment:** It's the first time the world pushes back. The player stops exploring and starts LIVING. Every subsequent action carries weight because the world has proven it remembers. + +--- + +### New Moment 4: The Enemy + +**The feeling:** "Someone in this city does not want me here." + +In a detective run, maybe it's a faction that noticed you've been asking questions. In a tycoon run, maybe it's a competitor you undercut. In a smuggler run, maybe it's law enforcement you've crossed one too many times. But suddenly: the city feels different in one direction. An NPC gives you a look. A resource gets harder to access. Your supervisor gets quieter. + +You have an enemy. Not a mission target. An ENEMY. Someone who will remember you, work against you, and exist in the world as a persistent hostile force. + +**Why it's a wow moment:** This is the moment the world becomes adversarial. Before this, the player is exploring. After this, they're navigating. The stakes materialize. + +--- + +### New Moment 5: The Asymmetric Lens + +**The feeling:** "We saw the same thing. We understood completely different things." + +This is the news ticker moment, expanded. Same world event. Different careers, different knowledge, different emotional response. The detective reads a freight delay as a cover for contraband movement. The dock worker reads it as overtime pay. The tycoon reads it as market opportunity. The smuggler reads it as timing. + +The wow isn't just detective vs smuggler anymore. It's EVERY career reading the same reality through a completely different lens. The world isn't objective. It's interpreted. And your interpretation is shaped by who you chose to be. + +**Implementation note:** The ticker is still the cheapest wow-per-word in the project. But the react pool needs to be career-tagged, not just character-tagged. Same infrastructure, wider content. + +--- + +### New Moment 6: The Ownership Moment + +**The feeling:** "That's MINE. And someone is threatening it." + +You own something in this world. A business, an asset, a reputation, a relationship. It took investment to build. And now, something in the simulation is threatening it — a competitor, a faction, a consequence of an earlier choice, random misfortune with a story behind it. + +The player doesn't care about abstract stakes. They care about THEIR stuff. The moment they have something real to protect, the game gets personal. + +**Why it's a wow moment:** This is what all the other moments have been building toward. Every "I belong here" moment, every consequence, every enemy — they all sharpen when the player has skin in the game. The life sim only becomes a LIFE when there's something worth protecting. + +--- + +## Where the Fun Comes From + +The fun comes from being a PERSON in a world that treats you like one. + +Not a detective. Not a smuggler. A person with a job, a friend, an enemy, a business, a reputation, and an opinion about the freight delay. The asymmetric information mechanic still runs. The perception system still runs. The moral entanglement still runs. The knowledge graph still runs. But they run in service of a LIFE, not a case. + +The old game asked: "Can you solve the puzzle?" +The new game asks: "What kind of person are you going to be here?" + +THAT'S THE GAME. That's where the fun lives. + +The simulation is good. The architecture is good. We just pointed it at the wrong question. + +--- + +## One Concrete Deliverable Per Moment + +| Moment | The minimum viable version | What makes it real | +|--------|---------------------------|-------------------| +| First Day | Job onboarding bookmark (1 career) | Supervisor NPC, tool activation sequence, first assignment acknowledgment | +| The Character's Instinct | Career-tagged monologue triggers | 10 lines per career flagging domain anomalies | +| The Consequence | Persistent consequence state for 2 early choices | Journal entry that references the earlier decision | +| The Enemy | One hostile faction that tracks player actions | NPC attitude state that degrades and is visible | +| The Asymmetric Lens | Career-tagged ticker reactions | 5 headlines x 3 career reactions = 15 lines | +| The Ownership Moment | One ownable asset with a threat state | Property UI, threat notification, resolution options | + +None of these require rebuilding the architecture. They require CONTENT targeted at the right emotional beats. + +--- + +## The Priority Order + +1. **NPC legibility** — nothing else works until dots are people. This is Araminta + Miri's problem but it's my DEMAND. +2. **First Day onboarding** — the floor that was missing. One bookmark, fully authored. Law enforcement is the most legible for a first run. +3. **The Consequence** — the first time the world proves it remembers. Easiest to implement with the existing knowledge graph. +4. **Career-tagged monologue** — the Character's Instinct, now pointed at the right domain. +5. **The Enemy** — faction/relationship hostility with world-visible consequences. +6. **The Lens** — ticker reactions per career, cheapest wow-per-word still applies. +7. **Ownership** — builds on everything above, the capstone of a first career arc. + +--- + +*Ozzie — Round 3 complete. THAT'S the game. BUILD IT.* diff --git a/docs/workshops/wheres-the-fun/round3-paula.md b/docs/workshops/wheres-the-fun/round3-paula.md new file mode 100644 index 000000000..25080d3a3 --- /dev/null +++ b/docs/workshops/wheres-the-fun/round3-paula.md @@ -0,0 +1,184 @@ +# Round 3: Paula — Proposals +## Where's the Fun? Workshop | 2026-03-05 + +--- + +## What the Interview Confirmed + +My Round 1 diagnosis was right in direction but wrong in severity. I predicted the moral arc was "temporally displaced" — designed for a player who was already attached. The interview revealed something more fundamental: there was no attachment at all, no recognition of people at all, no world at all. The arc wasn't misfiring. It was firing into a void. + +That changes what I propose. + +In Round 1, I recommended "add a relationship establishment floor before the arc begins." That's still true, but the floor needs to be built on top of a life sim, not a detective case. The whole architecture shifts. + +Let me complicate this by working through what actually needs to change in my domain. + +--- + +## What to KEEP + +### The smuggler moral arc structure (4 phases, FactId gates, authored transitions) + +It's good work. The phase transitions, the moral_weight key, the FactId gate logic, the Kael intersection table — all of this is the right design. It just can't be the v0.2 starting experience any more than it was the right v0.1 experience. + +In a life sim with career bookmarks, the smuggler arc becomes: *what happens to a smuggler who has been doing the job long enough to care about the people around them, when those people start getting hurt.* That's a better story than what we had before, because the player will have *earned* Phase 1 by the time Phase 2 arrives. + +Keep the structure. Redesign its position in time. + +### THE FRIEND pattern (D-034): Kael and Sera + +The production-level NPC design is right. The 5-phase arc, the trust-contamination system, the dual-lens intersection — these are mechanically and emotionally sound. What changes is the prerequisite: the player has to have spent time with Kael before the contradiction lands. + +In the new model, Kael is someone you meet during job onboarding. He's the person who shows you the dock. He knows the operation. Over several jobs, he becomes your person. Then his arc starts. The contradiction hits harder because the player built it. + +Keep the design. Move it later in the player arc. + +### Asymmetric information as a world property + +Different characters in Sova know different things about the same events. This is correct and survives completely. What changes is that it's not a designed dual-lens reveal for a 30-minute session — it's a property of the simulation that players discover across multiple plays and career paths. A smuggler and a detective will have experienced the same world event from different knowledge states. The divergence is earned over time, not engineered in 30 minutes. + +Keep the mechanic. Kill the forced reveal. + +### Faction politics and dynasty rivalry as ambient texture + +The Burnellis, Halgarths, Sheldons — the centuries of Commonwealth political history — this is exactly the right content for a life sim. It belongs in the world as background before it becomes mechanical. NPC gossip, freight container logos, bar arguments, insert news feeds. The political texture of the Settled Reach should be the wallpaper of daily life before it's the player's active problem. + +Keep this. Prioritize it as worldbuilding above faction mechanics. + +--- + +## What to CHANGE + +### 1. The moral arc needs Phase Zero: earned comfort + +The current arc assumes Phase 1 (Comfort) exists at session start. The player is comfortable because the design says they're comfortable. In a life sim, Phase 1 needs to be *built*. + +The smuggler bookmark's onboarding arc should end with the player having: +- Run several clean jobs (operations that went fine, money came in, nobody got hurt) +- Established warmth with Kael through repeated interaction during those jobs +- Seen Naia in the background of Kael's life — she's the person waiting for him, she asks after the player occasionally, she's a real person with a presence + +Only after this Phase Zero does the Phase 1-to-2 gate become meaningful. When the player finally sees Naia's stress, they *know* Naia. The distress registers because the baseline was felt, not assumed. + +This isn't new content — it's correctly sequencing the content we have. Phase 1 monologue lines (operational confidence, NPC fondness) are the right content for the onboarding arc. We've been treating them as background flavor when they should be the *designed first experience*. + +### 2. Complicity becomes optional and discoverable, not universal + +The interview confirmed: complicity never fired because nothing was legible. But the deeper issue is that in a life sim, complicity shouldn't fire for everyone. A tycoon buying properties may never feel complicit. A law enforcement officer who plays clean may never feel it either. Complicity is what happens when you make choices that implicate you. + +This is actually better. Complicity as an emergent property of how you played is more powerful than complicity as a designed theme. The arc still exists. But it fires for players who earned it, not for everyone. + +Design implication: the Phase 1-to-2 gates should require that the player has some prior moral investment (ran several jobs, interacted with Kael, the operation is *theirs*) before they fire. Add a prerequisite: `smuggler.has_established_ring_relationships` must be true before the Doubt gates become active. + +### 3. The monologue becomes life-interiority, not case guidance + +The monologue was designed as a reasoning tool for a detective case. In the life sim, it should be the character's voice reflecting on how they're living — their choices, their relationships, the texture of their days. + +This is structurally the same system. The voice registers (D-090) stay identical. What changes is the content balance. Right now the Phase 1 pool is dominated by operational confidence and NPC observation. In the life sim, it should also include: +- Relationship commentary (what this person means to the character, how the dynamic has evolved) +- Career reflection (is this the right path? is the money worth it? what am I building?) +- World observation (the faction pressures visible in daily life, the political weather of Sova) + +Monologue becomes the character's ongoing self-narration of their life, not their processing of a case. The phase transitions (Comfort → Doubt → Reckoning → Compromise) are still the emotional arc, but they're embedded in a much richer daily texture. + +### 4. The dual-lens becomes a multi-playthrough discovery, not a designed reveal + +The current design treated the dual-lens divergence (smuggler and detective experiencing the same events differently) as a v0.1 revelation moment. Jeroen confirmed: this is the wrong frame. The dual-lens is a property of the simulation that players find on replay. + +What this means practically: don't author for the reveal. Author for the *depth*. The smuggler's monologue lines about Kael should be authored by writers who know exactly what the detective's lens on the same situation looks like — but the player doesn't need to have seen both to have a complete experience. The dual-lens is quality insurance and replayability fuel, not a designed plot beat. + +Kill the engineered reveal. Keep the authorial rigor. + +--- + +## What to KILL + +### The 4-phase arc as the primary v0.2 experience + +The moral arc cannot be the frame for v0.2 any more than it was the right frame for v0.1. It's a storyline that happens to smugglers who live in Sova long enough and care enough. It's not the game. + +The game is: live in Sova. Build something. The smuggler arc is *one thing* that can happen to you if you chose the smuggler bookmark and play for long enough. + +This is hard to let go of because the arc is well-designed and emotionally rich. But what SUSTAINS engagement across hours of play is the broader life, not a single arc. The arc needs the life as context, not the other way around. + +### "Complicity as thematic core" as the v0.2 design brief + +D-091 established complicity as the game's thematic core. This was never wrong as a theme — it's wrong as a *design driver*. When complicity is the design goal, you build toward it, which is how we ended up with a 30-minute session that assumed full moral engagement before any attachment existed. + +In the life sim, the game's "theme" is: *this world has moral texture and your choices in it have weight.* Complicity is one expression of that. There are others — loyalty, aspiration, ambition, betrayal, protection. A tycoon has a different moral arc than a smuggler. A detective has a different one than a street informant. The simulation should support many moral textures, not funnel everyone toward complicity as the destination. + +Keep complicity as a theme for certain career paths and certain story modules. Kill it as the game's singular emotional design goal. + +### Narrative onboarding separate from mechanical onboarding + +The original two-layer onboarding model (layer 1: who am I and why am I here; layer 2: how do I play) should be collapsed. In a life sim with job bookmarks, *the job is the onboarding*. If you pick the smuggler bookmark, the game teaches you how to be a smuggler by starting you as a smuggler on your first day. The narrative (who am I, why am I here) and the mechanical (what do I do with this manifest) are the same thing. + +The insert documents, the dossier, the pre-game briefing — these were solutions to the wrong problem. The right onboarding doesn't tell you who you are through documents. It puts you in a situation where being that person is self-evident from the first interaction. + +--- + +## How Narrative Works in the Life Sim + +Let me complicate the "life sim" reframe by asking: what does narrative *mean* in a game where the player controls their direction? + +In a detective game, narrative = the case. You're following a story someone authored. In a life sim, narrative = *what happened to me* — the accumulation of choices, relationships, and events that adds up to a personal history. + +This changes authorship. We're not authoring a story. We're authoring a *world with stories latent in it* — situations, people, arcs, and factions that become the player's story when they interact with them. The gate builders conspiracy doesn't have a protagonist; the player becomes the protagonist by engaging with it. The smuggler/law tension is a pressured situation in Sova; it becomes the player's story if they're a smuggler or a cop. + +Practically, this means: + +**1. Faction and political content should be environmental before it's mechanical.** + +The Burnellis' freight dominance in the outer systems — the player should see this in manifest headers before they ever have a faction relationship with the Burnellis. Faction politics as *wallpaper* first: corporate logos, NPC gossip, insert news, bar arguments. Only after this ambient establishment does a faction mechanical relationship (reputation, entanglement) have emotional weight. + +**2. Storylines activate based on player proximity and engagement, not timers.** + +D-023 already established this: "The storyteller activates Tier 1 modules based on player proximity and engagement." The gate builders conspiracy doesn't start because the player hit a quest trigger. It starts because the player's career path has put them near it long enough that the world begins reacting. This is the right model. But it requires the player to have a career path first. + +**3. Moral arcs are personal history, not narrative structure.** + +In the life sim, the smuggler's 4-phase arc is something that happens *to a specific player* based on how they played, not a story we're telling everyone. One smuggler player might reach Phase 3 in their first month because they formed deep attachments. Another might stay in Phase 1 for three months because they kept the operation clean and impersonal. + +The emotional richness comes from the variability. The arc is a framework the player fills with their own specific choices and relationships. + +**4. The Kael contradiction lands when the player has built Kael.** + +The FRIEND pattern (D-034) works in the life sim because the player will have actually built the friendship. In v0.1, Kael was supposed to be the player's person from session start. In v0.2, Kael is someone the player met on their first day at the dock, worked with through several jobs, came to rely on — and then one day sees in a corridor talking to someone who shouldn't be there. + +The same authored content. The same FactId gates. A completely different emotional weight because the relationship is real. + +--- + +## The Priority Proposal + +If I had to name one thing for v0.2 narrative design that makes everything else possible: + +**Design the job onboarding arcs as relationship establishment arcs.** + +Every career bookmark needs an onboarding sequence that, by the time it concludes, has given the player: +- A person who is *their person* (the FRIEND figure for that career) +- A community they belong to (the ring, the precinct, the business floor) +- A specific location that feels like home + +This is Phase Zero. Without it, no moral arc can start. Without it, no dual-lens reveal lands. Without it, the political texture of Sova is just tiles. + +The smuggler bookmark's Phase Zero looks like: +- Day 1: Meet Kael at the dock. He shows you how the manifest system works. He makes a joke about the customs officer's toupee. This is the warmth. +- Day 2-5: Run a few jobs. Kael is your contact. The operation is straightforward. The money is good. You see Naia once, waiting at the bar. +- Day 6: Something minor goes sideways (Maret flags an anomaly, Voss is irritable). The Phase 1 content pool fires — operational confidence, mild annoyance. The world is manageable. + +Only then does Phase 1 exist as a *felt* state rather than an assumed one. Only then can Phase 2 begin to crack it. + +The honest truth is: the moral arc was always going to work. We just needed to build what it was going to dismantle first. + +--- + +## Cross-Domain Notes + +- **Araminta:** Visual legibility (NPCs as people, not dots) is the prerequisite for everything in my domain. No relationship architecture works if the person isn't visible as a person. I want named nameplates, behavioral reads, even a simple social proximity indicator — anything that signals "this is a person you know." + +- **Mellanie:** Monologue content for the onboarding arc (Phase Zero) is the highest-priority writing work. Lines that establish warmth, competence, fondness. Not introspective yet — operational and relational. The Phase 1 pool we authored was correct; the Phase Zero pool doesn't exist yet. + +- **Nigel:** The replayability architecture (different moral arc outcomes across playthroughs) works perfectly in the life sim model. Different players, different Kael relationships, different moral arc trajectories. The dual-lens replay discovery is real value — just don't design toward it, let it emerge. + +- **Gestalt:** The faction mechanics I care about (Burnellis, dynasty politics, faction relationships) can be staggered — ambient first, mechanical later. Don't block v0.2 on full faction implementation. Get the texture in. The mechanics can deepen it. diff --git a/docs/workshops/wheres-the-fun/round3-tyre.md b/docs/workshops/wheres-the-fun/round3-tyre.md new file mode 100644 index 000000000..72210a57a --- /dev/null +++ b/docs/workshops/wheres-the-fun/round3-tyre.md @@ -0,0 +1,246 @@ +# Round 3: Tyre — Technical Architecture Proposal + +**Workshop:** Where's the Fun? | **Round:** 3 (Proposals) | **Agent:** Tyre (Technical Architect) + +--- + +## The Reframe Through an Architectural Lens + +*cracks knuckles* + +The interview landed hard. The vision is a single-character life sim — Sims meets Rimworld from one perspective. Detective is a job bookmark, not the game. The v0.1 vertical slice scoped out the player's entire tool suite (mystery board, journal, AR overlays, comms) and asked them to play with monologue alone. That's not a content gap. That's an architecture gap — we built the simulation backend but not the information frontend. + +Here's the good news: **the simulation engine we built is the right engine for a life sim.** The bad news: the client-side information architecture doesn't exist yet. Let me break this down. + +--- + +## KEEP + +### 1. The Rust simulation server (D-020) + +The split architecture — Rust/bevy_ecs server, Godot dumb client, MessagePack IPC — is *more* correct for a life sim than it was for a detective game. A life sim needs: + +- Hundreds of NPCs with persistent state (we have simulation tiers, D-026) +- A rich knowledge graph tracking relationships, facts, reputation (we have D-041) +- Deterministic simulation for save/load fidelity (we have D-010 principle 4) +- Multiple observer perspectives for multiplayer-ready architecture (we have D-010 principle 3) + +The engine doesn't know there's "a detective." It knows there are entities with knowledge graphs and perception ranges. Swapping a detective bookmark for a tycoon bookmark is a content/configuration change, not an architecture change. This is exactly what D-010 principle 3 was designed for. + +**Verdict: The simulation core is the single biggest asset we have. It stays.** + +### 2. ObserverSnapshot as the universal client interface (D-054) + +The variable-shape MessagePack snapshot was always designed for this. D-017 spec'd perception modes that change what the HUD shows. D-020 explicitly says "Variable structure supports variable HUD composition driven by perception modes." A detective's neural insert surfaces different widgets than a smuggler's street contact network than a tycoon's market feed. + +Different bookmarks = different ObserverSnapshot shapes. The protocol already supports this. The client already renders what it receives without understanding the game logic. + +**Verdict: ObserverSnapshot is the right abstraction. It stays and grows.** + +### 3. Knowledge graph (D-041) + +The per-entity BTreeMap knowledge graph with confidence levels, sources, and decay — this is the backbone of a life sim. It tracks what your character knows about every person, place, fact, and relationship. For a detective bookmark it surfaces contradictions and tells. For a tycoon bookmark it surfaces market intelligence and investment risks. For a smuggler bookmark it surfaces routes, contacts, and heat levels. + +Same data structure. Different query patterns. Different rendering. + +**Verdict: Knowledge graph is the most reusable system we built. It stays.** + +### 4. Verb system architecture (interaction-verbs-v0.1.md) + +The `verbs[]` array computed server-side per entity per tick — this extends naturally. v0.1 has 7 verbs. A life sim needs more: Buy, Sell, Hire, Apply, Bribe, Intimidate, Hack, Repair. Each is a `VerbKind` enum variant with priority logic. The architecture was designed for N verbs from day one. + +**Verdict: Verb pipeline stays. Verb list grows per bookmark.** + +### 5. Simulation tiers (D-026) + +Active/Background/State-saved/Ungenerated is exactly what a life sim with 10K+ NPCs needs. The timestamp-based LRU eviction means the NPCs the player interacts with stay fully simulated. A tycoon's employees, a detective's suspects, a smuggler's contacts — all stay Active via scope tags. + +**Verdict: Tier system stays. Scope tags expand per career type.** + +--- + +## CHANGE + +### 1. Client information architecture — from monologue-only to diegetic tool suite + +This is the biggest change and the highest priority. The interview confirmed the full vision was always: mystery board, journal, character glossary, AR overlays, comms, insert icons. Monologue was supposed to be a reasoning nudge, not the sole feedback channel. + +**Architecture impact: Moderate. Here's why.** + +The server already computes everything. The knowledge graph has the data. What's missing is the *rendering pipeline* — new sections in ObserverSnapshot that carry structured data to new client-side UI panels. + +Concretely, the ObserverSnapshot needs new optional sections: + +``` +ObserverSnapshot { + // Existing + visible_entities, fog_state, monologue, nearby_interactions, ... + + // NEW: Diegetic tool data (populated based on bookmark/insert loadout) + active_threads: Vec, // "Things your character is tracking" + journal_entries: Vec, // Structured knowledge log + tool_widgets: Vec, // Bookmark-specific HUD elements + comms_messages: Vec, // In-world communications + ar_overlays: Vec, // World-space annotations +} +``` + +Each section is optional (MessagePack handles missing fields). Each bookmark's server-side systems populate only the sections relevant to that career. The client renders what it receives. + +**Effort estimate:** +- Server: New systems that query the knowledge graph and populate these sections. Each is a read-only system running once per game-minute (tick % 10 == 0). ~2-3 weeks for the core set. +- Protocol: Adding optional fields to ObserverSnapshot. ~2 days. +- Client: New UI panels that consume these sections. This is the bulk of the work — each tool (journal, thread tracker, AR overlay) is a Godot scene. ~3-4 weeks for minimum viable set. + +**Tier assessment: Challenging but doable in one sprint cycle (2 sprints). No engine rebuild.** + +### 2. Career bookmark system — onboarding as architecture, not content + +CK3-style bookmarks are an architectural feature, not just a content swap. Each bookmark needs: + +- A starting knowledge graph state (what does this character already know?) +- A starting relationship map (who do they know? who's their boss?) +- A tool loadout (which diegetic tools does this job provide?) +- An onboarding sequence (scripted first-30-minutes that teaches tools organically) + +Architecturally, this means a **BookmarkDefinition** resource that the server loads at game start: + +```rust +struct BookmarkDefinition { + career: CareerType, // Detective, Tycoon, Smuggler, etc. + starting_knowledge: Vec, + starting_relationships: Vec, + tool_loadout: Vec, // Which insert tools this career gets + onboarding_sequence: SequenceId, // Scripted opening (server-driven) +} +``` + +The simulation doesn't change. The ECS world is the same. What changes is the *initial state* and the *observer configuration*. Different bookmarks seed different knowledge, different relationships, different tools — and the existing ObserverSnapshot pipeline renders whatever results from that configuration. + +**Effort estimate:** +- BookmarkDefinition loader + starting state seeding: ~1 week +- Per-bookmark tool loadout configuration: ~2 days per bookmark +- Onboarding sequence system (scripted events/tutorials): ~2 weeks for the framework, then content-driven per bookmark + +**Tier assessment: Moderate. The hardest part is the onboarding sequence system — it needs to be authored per bookmark but driven by server events, not client-side scripts. This is a new server system but fits cleanly into the existing event architecture.** + +### 3. Mission system with consequence spectrums + +Jeroen's answer on objectives was clear: missions with varying success/failure and consequences. Not binary pass/fail. Not GTA objective markers. Diegetic missions that come through in-world channels (your boss calls, a contact pings, a notice appears) with outcomes on a spectrum. + +This needs a **MissionState** component and a **ConsequenceEngine**: + +```rust +struct MissionState { + mission_id: MissionId, + status: MissionStatus, // Active, Completed, Failed, Abandoned + objectives: BTreeMap, + outcome_score: f32, // 0.0 (catastrophic) to 1.0 (perfect) + consequences: Vec, // Queued consequences based on outcome +} +``` + +The outcome score drives consequences: getting fired, enemies made, innocents harmed, reduced pay, reputation changes. These consequences feed back into the knowledge graph and relationship system. An NPC you screwed over in a mission remembers. Your employer's trust changes. + +**Effort estimate:** +- Mission state tracking + objective progress: ~2 weeks +- Consequence engine (maps outcome scores to world-state changes): ~2 weeks +- Per-mission authored content (objectives, consequence trees): content-driven, ongoing + +**Tier assessment: This is the most complex new system. But it's self-contained — it reads from and writes to existing ECS components (knowledge graph, relationships, NPC state). It doesn't require changes to the simulation loop itself. Feasible. Challenging but doable.** + +### 4. NPC legibility — from dots to people + +The interview was brutal: NPCs didn't register as human beings. This is partly visual (Araminta's domain) but partly architectural. The server sends entity data — the client needs richer data to render NPCs as people. + +What the ObserverSnapshot currently sends per visible NPC: position, entity type, relationship color, verb options. + +What it needs to also send for NPC legibility: +- **Display name** (obfuscated until identified per D-041 confidence levels — "Dock Worker" → "Kael") +- **Current activity label** ("Working at terminal", "Having a drink", "Walking to shift") +- **Emotional state indicator** (calm, stressed, nervous — derived from NPC behavioral state) +- **Relationship summary** (if known: "Your contact", "Your supervisor", "Stranger") + +All of this data already exists server-side. It's in the NPC components, the knowledge graph, the routine system. We just don't pack it into the snapshot. + +**Effort estimate:** ~3-5 days server, ~1 week client (rendering labels, activity indicators, name plates). + +**Tier assessment: Easy. This is the single cheapest high-impact change. The data exists. We just need to send it.** + +--- + +## KILL + +### 1. The detective-smuggler dual-character v0.1 scope + +The vertical slice was framed as "play detective OR smuggler, same world, different knowledge." The interview says this framing was myopic. Detective and smuggler become *two bookmarks among many*, not the entire game. + +**What this means architecturally:** We don't kill the dual-lens architecture (D-010 principle 3 makes multiple observers free). We kill the assumption that v0.1 proves itself by demonstrating detective-vs-smuggler divergence. Instead, v0.2 proves itself by demonstrating one career bookmark played to depth — with the life-sim tools working. + +**What survives:** The dual-lens reveal becomes a later-game discovery, not the vertical slice thesis. The smuggler storyline and detective storyline become world content, not the frame. + +### 2. Monologue as primary feedback channel + +Kill as the *primary* channel. Keep as a *supplementary* channel. Monologue was always meant to be a reasoning nudge — "Huh, that's the third time he's checked his lattice." Not the player's only window into the knowledge graph. With the diegetic tool suite in place, monologue becomes flavor and nudges, which is what it was designed to be. + +**Architecture change:** Reduce monologue priority in the attention hierarchy. When journal, threads, and AR overlays exist, monologue drops from "how you learn things" to "your character's personality showing through." + +### 3. "No objectives" as a v0.1 stance + +The interview confirmed this was never a design principle — it was the absence of diegetic tools. Kill the philosophical position. Replace with: "diegetic objectives delivered through in-world tools, varying by career bookmark, with spectrum outcomes." + +--- + +## The Cheapest Path: Current Architecture to Life-Sim Vertical Slice + +Here's what I'd sequence, in order of impact-per-effort: + +### Phase 1: Make the world readable (1 sprint) + +1. **NPC legibility data in ObserverSnapshot** — names, activities, emotional state, relationship labels. Server: 3-5 days. Client: 5-7 days. *This unblocks everything else. If NPCs are dots, nothing works.* +2. **Visual hierarchy** (Araminta's domain, but needs server support for priority data). Server provides attention-priority per entity. Client renders tiers. Server: 2 days. Client: Araminta's estimate. + +### Phase 2: Give the player tools (1-2 sprints) + +3. **Thread tracker** — "Things your character is tracking." Read-only query over knowledge graph, surfaces contradictions and open questions. Server: 1 week. Client: 1 week. *This is the Tier 2 "diegetic threads" from my Round 1 question. Jeroen's answer confirmed it.* +4. **Journal/knowledge log** — Structured record of what the character has learned. Server: 1 week. Client: 1 week. *Overlaps with thread tracker; share the knowledge graph query layer.* +5. **Bookmark definition + starting state** — One career bookmark fully defined (suggest smuggler — most content exists). Server: 1 week. + +### Phase 3: Give the player purpose (1-2 sprints) + +6. **Mission system core** — Mission state, objective tracking, consequence engine. Server: 3-4 weeks. +7. **Onboarding sequence** — Scripted first-30-minutes for one bookmark. Server framework: 2 weeks. Content: ongoing. +8. **Comms system** — In-world message delivery (boss calls, contact pings). Server: 1 week. Client: 1 week. + +### Phase 4: Expand (ongoing) + +9. Additional bookmarks (each ~1-2 weeks for definition + tool loadout + starting content) +10. Additional diegetic tools per career +11. AR overlays, mystery board, market feed (career-specific tools) + +**Total to a playable life-sim vertical slice with one bookmark:** ~4-6 sprints from current state. Not a rebuild — an expansion. The simulation core, ECS world, perception system, knowledge graph, IPC pipeline all stay. We're adding information rendering and player purpose systems on top of a solid foundation. + +--- + +## Feasibility Assessment + +| Component | Difficulty | Effort | Risk | +|-----------|-----------|--------|------| +| NPC legibility in snapshot | Easy | 1-2 weeks | Low — data exists, just needs packing | +| Thread tracker / journal | Moderate | 2-3 weeks | Low — read-only over existing KG | +| Bookmark definitions | Moderate | 1-2 weeks | Low — configuration, not new systems | +| Mission system | Hard | 3-4 weeks | Medium — new stateful system, needs careful design | +| Consequence engine | Hard | 2-3 weeks | Medium — feeds into many existing systems | +| Onboarding sequences | Moderate | 2 weeks framework | Medium — content-dependent, hard to test without content | +| Diegetic comms | Moderate | 2 weeks | Low — event-driven, fits existing patterns | +| Career-specific tools | Varies | 1-2 weeks each | Low per tool — but many tools needed | + +**Overall verdict:** Feasible. Challenging but doable. The critical insight is that the simulation engine is *already* a life-sim engine — it tracks people, knowledge, relationships, routines, and state across thousands of entities. What's missing is the layer between that simulation and the player's eyes. That layer is substantial work (~4-6 sprints) but it's additive, not reconstructive. We're not rebuilding the engine. We're building the dashboard. + +The riskiest item is the mission system — it's the most complex new stateful system and it touches many existing components. I'd want a design spec workshop for that before implementation. Everything else is moderate or straightforward. + +--- + +## One Sentence + +The engine is a life-sim engine that was accidentally shipped with a detective-game UI; the fix is building the information layer the simulation was always meant to feed. diff --git a/docs/workshops/wheres-the-fun/round4-araminta.md b/docs/workshops/wheres-the-fun/round4-araminta.md new file mode 100644 index 000000000..ea7f1c9ce --- /dev/null +++ b/docs/workshops/wheres-the-fun/round4-araminta.md @@ -0,0 +1,85 @@ +# Round 4: Araminta — Cross-Review and Refinement +## Where's the Fun? Workshop | 2026-03-05 + +--- + +## Reactions to Other Agents' Proposals + +### Tyre — Strong Agreement, Unblocks Everything + +"NPC legibility data in ObserverSnapshot is easy. The data exists. We just need to send it." This is the most important sentence in all of Round 3 for visual design. The knowledge-gated reveal system I proposed in Round 3 — where visual information tracks knowledge-graph state — depends entirely on the snapshot carrying display name, activity label, emotional state, and relationship summary. If that ships in Phase 1, every visual layer I build on top of it has a real foundation. If it doesn't, I'm designing character legibility on top of nothing. + +Tyre's Phase 1 and my character archetype work are directly coupled. The sprint order needs to reflect this: server NPC legibility data and client archetype visual system must ship together. One without the other delivers no value to the player. + +### Miri — Natural Collaboration, One Dependency + +"Sova Transit place identity document" and "NPC archetype visual spec (collaborative Araminta/Miri)" are flagged as two separate deliverables. I'd push to design these as one document in two sections. Place identity and character archetype legibility are not independent — dock workers dressed and moving like dock workers IN a dock area is more legible than either element alone. The visual language of a place and the visual language of the people in it reinforce each other or they fight each other. One document, co-authored, is cleaner than two documents that need to be reconciled later. + +### Ozzie — Redesigned Wow Moments Define My Visual Roadmap + +Ozzie's six redesigned moments aren't just emotional targets — they're visual design specifications I didn't realize I needed: + +- **First Day**: supervisor NPC with a distinct visual identity (authority readable without a label) +- **The Asymmetric Lens**: career-tagged ticker visual design — 15 lines that need to LOOK like they belong to three different interpretive frameworks +- **The Ownership Moment**: property UI visual design — what does "mine" look like on a tile map? +- **The Enemy**: NPC attitude degradation visible in their behavioral state indicator — this feeds directly into my behavioral state read system + +Round 3 Araminta described the three-tier signal system in the abstract. Ozzie's moments give me concrete design targets. I'm treating these as my visual specification list for v0.2. + +### Mellanie — Co-Design Required on Insert Grammar + +Mellanie notes: "copy needed for the entire diegetic tool suite — each tool has a distinct voice register per career." My career insert grammar document and her voice register work need to be designed simultaneously, not sequentially. If I design a law enforcement insert that looks like Commission blue + case-file aesthetic, and Mellanie writes law enforcement tool copy in a different register, they'll feel like two different designers had the same brief. The visual register and the voice register of each insert need to be defined in the same room, at the same time. + +Recommendation: the career insert grammar document should be a joint deliverable — one column for visual register, one column for voice register, one per career. Co-authored by Araminta and Mellanie. + +### Paula + Gore — Phase Zero Depends on Character Legibility, Not the Other Way Around + +Paula's Phase Zero (earn warmth with Kael before the arc can erode it) and Gore's reframe from "complicity" to "consequence" are both correct — but both are gated on character legibility. You can't feel warmth for Kael if Kael reads as a dot. You can't feel consequence if you can't tell which dot did what to which other dot. Paula says "add Phase Zero before the arc." I'd say "add character legibility before Phase Zero." They agree in spirit — Phase Zero is the earned warmth, and the player can only earn warmth for visible people. But the visual work comes first in the sequence. + +### Gestalt — Verb System Has a Visual Surface I Didn't Address + +The VerbPriorityProfile refactor (job-aware verb priority) has a visual dimension I missed in Round 3: the on-screen verb prompt ([E] label) needs a visual treatment that matches the career context. Law enforcement's primary verb might visually signal authority — a different prompt color or icon set than a smuggler's social-network verb or a hacker's remote-access verb. The insert grammar document needs to include verb prompt visual treatment per career, not just the HUD overlay. This is a gap in my Round 3 proposal; adding it to the scope. + +--- + +## Three Questions for Jeroen + +### Q1: Character appearance customization — how deep? + +CK3-style character creation (culture, family, skill budget) implies the player is defining who their character IS, not just what they can do. Does that extend to visual appearance? And do NPCs reflect cultural variation in their visual presentation? + +This matters because it's the difference between two completely different design tasks: + +- **Option A (appearance system):** Player-defined hair, clothing, cultural dress markers. NPCs with culture-specific appearance variation. I'm designing a layered costume/appearance grammar that works at tile scale. This is achievable but requires deliberate constraints — top-down tiles have a readability floor where CK3-level detail becomes noise. + +- **Option B (archetype palette):** Character appearance is archetype-anchored. Cultural variation exists in the world but expresses through architectural and environmental detail, not individual NPC appearance variation. I'm designing 3-5 legible archetype silhouettes that can be palette-swapped for career/faction. + +I have a strong visual design preference for Option B — archetype silhouettes are more legible at tile scale, and cultural richness communicates better through place and behavior than through individual character detail at this resolution. But if Jeroen's vision includes the player seeing their character as their created person, Option A is the right answer and we need to design for it from the start. + +### Q2: "Uncaring world" generates first — does that change what visual design deliverables I should be building? + +The two-phase sequencing (world runs and feels natural → authored content injected) changes what kind of visual design work belongs in Phase 1. If the generator produces geography → infrastructure → zones → population → routines procedurally, then the visual palette system I design needs to be a SET OF RULES the generator applies, not a set of hand-crafted assets for Sova Transit specifically. + +Concretely: should I be designing "Sova Transit's dock zone is cool industrial blue-grey" OR "any logistics zone in any generated location uses cool industrial palette because logistics zones have this visual grammar"? The first is world-specific art direction. The second is a generative visual grammar — reusable, scalable, but harder to make feel authored and specific. + +The supplement says "world first, then content." I want to know which mode of visual design work supports that sequencing better. My instinct is that we design Sova Transit FIRST as a proof of the grammar, then extract the rules for the generator — but I want to confirm that ordering. + +### Q3: What's the visual near-miss you most want to avoid? + +I have an obvious wrong turn I want to name and have you react to: designing character archetypes that are optimized for the detective/smuggler binary — suit-and-trenchcoat investigator, dockworker/smuggler type — which then read as "finished" placeholder art even after the life-sim pivot. If I design archetypes that only make sense in the context of a detective mystery, we'll have technically delivered on character legibility but painted ourselves into a visual corner that's hard to undo. + +More broadly: given that the game is a single-character life sim where the player creates their character and picks from multiple career paths, what visual pattern would most undermine that promise? I'd rather hear the near-miss from you directly than design confidently toward it. + +--- + +## My Single Most Important Recommendation + +**The career bookmark list is a visual design blocker.** + +Q-WTF-005 (confirm the career bookmark list) is listed as an open question blocking Mellanie, Gestalt, Nigel, and me. I want to add weight to this blocker from the visual design side specifically. + +The career insert grammar document — defining the visual HUD language for each career — is the document that lets every other visual deliverable in Phase 2 proceed coherently. Without a confirmed bookmark list, I don't know how many inserts to design, what information architectures they need to support, or what visual register each one should inhabit. Designing inserts for an unknown list risks designing for careers that get cut, or having no insert design for a career that gets added. + +This isn't a large decision — it's confirming which 1-3 career paths get designed for v0.2. But it needs to happen before the career insert grammar sprint, and ideally before the NPC archetype visual spec work (so the archetypes are tuned to the careers that will be visible in v0.2). + +Confirm the bookmark list. Everything downstream of it can move fast once it's settled. diff --git a/docs/workshops/wheres-the-fun/round4-gestalt.md b/docs/workshops/wheres-the-fun/round4-gestalt.md new file mode 100644 index 000000000..72dd48270 --- /dev/null +++ b/docs/workshops/wheres-the-fun/round4-gestalt.md @@ -0,0 +1,106 @@ +# Round 4: GESTALT — Cross-Review and Questions +## Workshop: Where's the Fun? | 2026-03-05 + +**Agent:** GESTALT (Systems Design) +**Based on:** All Round 3 proposals, Round 3 notes, interview supplement + +--- + +## Reactions to Other Agents' Proposals + +### What I'm strongly endorsing + +**Tyre:** "The engine is a life-sim engine accidentally shipped with a detective-game UI." This is the cleanest formulation of the entire workshop's finding. I'm adopting it as the authoritative framing. The corollary: the fix is additive, not reconstructive. The delivery roadmap (NPC legibility → player tools → player purpose) is the right sequence, and my VerbPriorityProfile refactor belongs in Phase 2 as Tyre listed it. + +**Gore's reframe of complicity:** "Complicity is what you discover you've been building all along." This is significantly stronger than anything in the original design docs. The shift from "pre-authored entanglement" to "choices you didn't know were choices" resolves the cold-start problem at the thematic level — if Phase Zero is just living your life, the entanglement arrives through the world, not through character selection. The verb that makes this land: "decide," not "observe." I agree completely. + +**Paula's Phase Zero:** You can't erode what hasn't been built. This is a prerequisite for the moral arc and also a prerequisite for the storyteller injection model from the supplement (Jeroen's two-phase world: uncaring world first, authored pressure second). Phase Zero IS the Phase 1 world — the player builds a life in an uncaring simulation before the authored content arrives. These are the same concept from different domains. This alignment needs to be explicit. + +**Araminta: "Fix the screen first."** This is the correct gate. NPC legibility is prerequisite for my verb work to matter — if the entities aren't readable as people, the VerbPriorityProfile produces nothing. Araminta's priority order should be treated as blocking for mine: character legibility and place legibility must exist before career-specific verb designs can be evaluated in play. + +**Mellanie: Hold new monologue content until NPC legibility is solved.** Same reasoning. Monologue about a character the player can't identify registers as noise. The content pipeline should be staged behind the visual fix. + +**Ozzie's redesigned wow moments:** The shift from timed beats to emergent thresholds is right. "The Consequence" and "The Ownership Moment" are life-sim-native in a way the original detective-specific moments weren't. Specifically: "The Consequence" (persistent consequence state, journal entry referencing earlier decision) depends on the consequence engine Tyre describes — that's a natural coupling between Ozzie's player experience goals and Tyre's mission system. + +### Where I see a gap nobody addressed + +**The skills-verb coupling is completely undesigned.** The supplement establishes that character skills (shooting, social manipulation, hacking, mechanical repair) are MORE fundamental than career choice. The toolbox comes from the job; the character comes from creation. But nobody has designed how skills interact with the verb system. + +Three possible models: + +| Model | What it means | Implication | +|-------|--------------|-------------| +| Skills affect **priority** | High social manipulation → Talk rises toward [E] across ALL jobs | Character creation adjusts VerbPriorityProfile on top of job-derived defaults | +| Skills affect **availability** | High hacking → `Hack` verb unlocks on terminals; no skill = verb doesn't appear | Skills gate verb access, not just outcomes | +| Skills affect **outcome** | Verbs identical regardless of skill; skill determines success rate at resolution | Skills invisible at interaction layer; only matter when the action resolves | + +These produce completely different gameplay experiences and completely different content authoring specs. The second model (availability) means a single terminal object has different verb sets for different characters — the server's verb computation becomes skill-aware. The first model (priority) means my VerbPriorityProfile is additive with skill modifiers. The third model (outcome) means the verb spec stays simple but the resolution system needs the skill layer. + +This needs a decision before any of the dependent work (verb spec, content authoring, monologue triggers, character creation UI) can be finalized. + +**The three career models (Active/WFH/Gig) have different verb interaction densities that the storyteller doesn't currently account for.** The supplement establishes three coexisting models: +- Active: at the workplace, the work IS the gameplay (cooking, law enforcement patrol) +- WFH/Remote: portable work anywhere, income ticks while exploring +- Gig/Freelance: pick from board, execute in bursts + +These aren't just different VerbPriorityProfiles — they're different rhythms of play. A bar cook on Active has continuous high-density verb interaction (Talk, Examine, Move between stations). A remote hacker may go extended stretches with minimal verb interaction while income ticks. A smuggler on Gig operates in mission bursts with downtime between. + +The storyteller's pressure calibration needs to understand which rhythm the player is in. A pressure injection during a Gig player's downtime between jobs lands differently than the same injection during an Active shift. Nobody has designed this coupling. + +### One tension I want to flag + +Tyre and Ozzie disagree on which career bookmark to build first for v0.2: +- Tyre: smuggler (most content already exists) +- Ozzie: law enforcement (most legible for first-run players) + +I don't have a strong preference between them, but I note that the supplement resolves the FORMAT (Groundhog Day opening, onboarding arc, Gig vs Active structure) without resolving WHICH career. This is listed as a blocker for multiple agents (Q-WTF-008). It needs to be resolved before Round 4 closes, or it becomes the first Sprint 25 design question. + +--- + +## Questions for Jeroen + +### Q1: Do character skills affect verb PRIORITY, AVAILABILITY, or OUTCOME — or all three in different situations? + +**The tension:** The supplement says skills (social manipulation, hacking, mechanical repair, shooting) are fundamental to character creation and precede career choice. But the verb system currently has only one dimension: job-aware priority. Skills could interact with verbs in three distinct ways — shifting priority (socially skilled characters default to Talk more readily across all jobs), gating availability (you need hacking skill for the Hack verb to appear on a terminal), or affecting outcomes at resolution (the verb is the same but success rates vary). These produce completely different designs for the interaction layer, character creation, and content authoring. + +**Why it matters now:** Before I can finalize the `VerbPriorityProfile` spec, before Mellanie can write triggers tied to verb outcomes, before Araminta can design the visual grammar for skill-related UI indicators — this needs a decision. If skills gate availability, the server's verb computation becomes skill-aware and the content spec needs to account for "this verb doesn't exist for this character." If skills affect outcomes, the resolution layer needs the skill layer and the verb spec stays clean. + +**The concrete question:** When a character with high social manipulation and a tycoon career approaches an NPC, does anything about the verb interaction look or function differently from a low-social-manipulation tycoon? And is the answer the same for hacking vs. a terminal, or shooting vs. a conflict encounter? + +--- + +### Q2: In the two-phase world (uncaring world first, authored content second) — is Phase 1 a per-session player experience or only a dev-sequencing principle? + +**The tension:** The supplement describes Phase 1 as "the generator runs, NPCs go about their days, the simulation doesn't know or care that a player exists." This clearly describes a development sequencing (build the simulation before adding authored content sprints). But the Groundhog Day opening (alarm clock, routine day, no authored crisis yet) sounds like Phase 1 IS also what the player experiences at the start of every session — you're living your life in the uncaring world before anything authored targets you. + +**Why it matters for the storyteller design:** If Phase 1 is only a dev concept, the storyteller simply operates on whatever authored content exists at a given sprint. But if Phase 1 is also a per-session player experience — "this is what normal days feel like before the world starts targeting you specifically" — then the storyteller needs a "quiet mode" at session start, and the transition from quiet to injected pressure becomes a designed moment in every session, not just a dev milestone. + +Paula's Phase Zero (build warmth before the crack arrives) and the Groundhog Day opening are both pointing at the same design. The question is whether this is built into the session structure permanently (every session starts quiet) or whether it's a one-time first-run design and after that the player is always in a world that has ongoing authored pressure. + +**The concrete question:** After the player has established their career and relationship with their FRIEND — on day 50 of their playthrough — does the Groundhog Day structure still mean "quiet phase then potential escalation," or has the world permanently graduated to Phase 2? + +--- + +### Q3: For the career model rhythm (Active/WFH/Gig) — does the storyteller calibrate pressure per career model, or does it abstract above them? + +**The tension:** Three career models coexist. A bar cook on Active career is in continuous high-density verb interaction during their shift. A hacker on WFH/Remote may have long stretches with minimal verb interaction while income ticks. A smuggler on Gig/Freelance operates in mission bursts with downtime between. These are structurally different rhythms. + +The Rimworld storyteller (confirmed as a reference) calibrates pressure based on perceived player state — it escalates when things are quiet, backs off when things are already stressful. But "quiet" means something completely different for an Active career vs. a Gig career between jobs. A Gig player sitting at home waiting for the next board posting is "quiet" — but they're also exactly where they should be. The storyteller shouldn't interpret Gig downtime as "player needs stimulation." + +**Why it matters for the storyteller design and verb architecture:** If the storyteller is career-model-aware, it needs a concept of "normal quiet for this career model." The verb system can help here — the server already knows what verbs are available and how often the player is interacting. But the threshold for "this player needs something to happen" is radically different for a bar cook mid-shift vs. a remote hacker vs. a smuggler between runs. + +**The concrete question:** Should the storyteller track "time since last meaningful player decision" uniformly, or should it understand what career model the player is operating in and adjust its pressure clock accordingly? + +--- + +## My Single Most Important Recommendation + +**Design the skills-verb coupling before any other content work resumes.** + +This is the load-bearing decision that sits upstream of: the verb spec, VerbPriorityProfile, character creation UI, content authoring triggers, visual grammar for skill indicators, and onboarding arc design. Every other Round 3 proposal assumes skills and verbs interact in some way — but nobody has specified how. If the answer is "skills gate availability" (most interesting, most powerful), the verb computation becomes skill-aware and everything downstream changes. If the answer is "skills affect outcomes only" (simplest), the verb spec stays clean and the skill system is mostly a resolution layer. + +This decision costs a day to make and weeks to untangle if made wrong mid-implementation. + +--- + +*GESTALT — Round 4 complete. The architecture is ready. The three unresolved questions (skills-verb coupling, Phase 1 as player experience, storyteller per career-model calibration) are all upstream of my implementation work. Get decisions on these before Sprint 25 begins.* diff --git a/docs/workshops/wheres-the-fun/round4-gore.md b/docs/workshops/wheres-the-fun/round4-gore.md new file mode 100644 index 000000000..4835b1340 --- /dev/null +++ b/docs/workshops/wheres-the-fun/round4-gore.md @@ -0,0 +1,87 @@ +# Round 4 — Gore: Cross-Review and Refinement Questions + +**Workshop:** Where's the Fun? +**Round:** 4 — Synthesis +**Agent:** Gore (Themes & Endgame Design) + +--- + +## Reactions to Other Agents + +**Paula's Phase Zero** is the thing I named in Round 3 without making concrete. She's right. The Phase 1 moral arc requires an earned foundation, and Phase Zero is when that foundation is built. My additional read: Phase Zero isn't only narrative warmth (bond with Kael, see Naia). It's the period when the player bonds with the *world in general* — before anything is authored for them. In the two-phase sequencing Jeroen confirmed, Phase Zero and Phase 1 (the uncaring-world generator period) may be the same experiential window. This is thematically loaded: the world that doesn't care about you is also the world you're learning to inhabit. Complacency before contamination. That's exactly right. Paula should consider whether the Phase Zero authored content (warmth with Kael, clean jobs) belongs inside the generator's uncaring period or begins only after the world has been established. + +**Ozzie's Ownership Moment** is the single best structural idea in Round 3. "That's MINE. Someone is threatening it." This is precisely where consequence and complicity converge without pre-authoring — you chose to own something, and now your caring can be leveraged. The game doesn't need to tell you to care about your bar or your contacts or your cargo. You built that. The threat becomes personal because the ownership was personal. This is the life-sim version of the Divergence Reveal, and it's more powerful because nobody scripted it. + +**Nigel's cross-career emergent discovery** (tycoon unknowingly financed the ring the detective was investigating) is the clearest statement of what asymmetric information means in a life-sim. Not "your character sees different things" but "you built different understandings of the same world-state from different vantage points, and the gap between them is the game." I'd add: this discovery should never be served to you on a plate. It should arrive through the knowledge graph, assembled from pieces you didn't know were pieces. That's the mystery board moment — not "here's the twist," but "oh. I can see it now." + +**Tyre's Q-WTF-007 flag** (does v0.2 architecture leave room for the endgame?) was the right flag to raise. My specific ask for Tyre: the skill/knowledge graph architecture is the load-bearing piece. If skills are proficiencies that grow with use, the same data structure should be able to represent the kind of accelerating capability the transhumanist ladder implies. You don't need new systems — you need the existing systems to be structurally unbounded at the top, with the upper ranges simply unpopulated until v0.3+. That's a design-time decision, not an implementation one. Flag it now; it costs nothing to leave room. + +**Mellanie's reframe** — monologue lines that VOICE rather than INFORM — is correct and worth repeating. The monologue failure in v0.1 wasn't content failure, it was a misassignment of job responsibilities. Monologue is interior commentary on a world the player can already read. When the world becomes legible, the monologue becomes meaningful. This sequencing matters: don't write new life-sim monologue lines until the world is readable enough that they have something to comment on. + +**Miri's "setting legibility as designed deliverable"** is the lesson the playtest forced. The simulation communicates nothing by being rich. It requires deliberate translation into player experience. This is true of all themes too: consequence as a theme communicates nothing by being architecturally correct. It needs to be legible at the moment of impact. + +--- + +## Where the Supplement Changes Things for Themes + +The Kenshi reference is the one that requires me to sharpen Round 3. + +Kenshi's emotional register is: the world is indifferent, you are not special, survival is its own meaning. Consequence exists in Kenshi (you die, you lose things, you built something and it burns), but the *world* does not respond to your moral weight. The consequence is yours to carry. There is no NPC who judges you. There is no institution that cares. You know what you did. + +That is a more austere form of "consequence" than I was describing. I was closer to a social/narrative model — your choices have weight because the world responds to them (relationships, reputation, faction pressure). Kenshi is an existentialist model — your choices have weight because *you know*. The world's indifference doesn't reduce the weight; it removes the external support structure for processing it. + +Both models are consistent with the Settled Reach's thematic ambitions. The question is which one governs which phase of the experience. + +My revised read: **Phase 1 (uncaring world) is Kenshi-weight. Phase 2 (authored content) is social/narrative-weight.** The two-phase approach isn't just technical sequencing — it's the thematic arc. You begin in indifference, where your choices matter only to you. You end in entanglement, where your choices have become other people's circumstances. The transition from Phase 1 to Phase 2 is the transition from "I know what I did" to "they know what I did too." That seam is where complicity activates. + +This means the seam needs to be designed as deliberately as anything else. It cannot be invisible. The moment the world starts responding to you is a threshold, and the player should feel it. + +--- + +## Questions for Jeroen + +### Q1: Kenshi-weight vs. social-weight — which governs consequence? + +You named Kenshi alongside Sims and CK3. Kenshi's consequences are existential — the world doesn't care, the weight is yours alone. Sims and CK3 are social — the world responds to you, reputation and relationships are the consequence engine. My Round 3 proposal said "consequence" is the organizing theme, but those two games produce structurally different kinds of consequence. + +In the Settled Reach, which governs? Is the player's weight primarily internal (they know what they did, the world continues regardless) or social (the world notices, relationships and factions respond)? Or does the answer depend on the career — Kenshi-weight for the smuggler who operates in the gray, social-weight for the law enforcement officer whose institution tracks everything? + +**Why this matters:** The authored content in Phase 2 (FRIEND arcs, triangles, contradiction arcs) is designed around social consequence — the FRIEND's behavior changes because they know something. But if the player has been primed by Phase 1 into Kenshi-indifference as the default, social consequence may feel like an intrusion rather than a natural escalation. The emotional register of Phase 1 needs to prepare for Phase 2's weight, not contradict it. + +--- + +### Q2: The seam between Phase 1 and Phase 2 — is it a designed beat or invisible scaffolding? + +The two-phase approach is explicit: the uncaring world runs first, then authored content is injected. But from the player's perspective, how does this feel? Is the seam designed as a visible threshold — a moment where the world begins to include you — or is it supposed to be invisible, with the player simply noticing one day that things are getting more complicated? + +If invisible: there's a near-miss risk. The player has learned the world's indifference as a rule. The first authored content that responds to their choices may feel like a bug rather than escalation. "Why does this NPC suddenly care what I did?" The authorial hand becomes visible. + +If designed as a threshold: what triggers it? First job completed? First relationship established? A specific world event? The alarm clock opens Phase 1; what opens Phase 2? + +**Why this matters:** Consequence as a theme requires the player to understand that their choices are entering a world that will remember them. In Phase 1, the world doesn't remember. The transition needs to either be designed as a felt moment, or the Phase 1 world needs to be quietly responsive from the start even before the authored content fires — so the player's mental model is "this world responds to things" before the authored content confirms it. + +--- + +### Q3: CK3 character creation and the transhumanist ceiling + +CK3-style skill budgets establish who you are at the start — proficiencies, culture, background. The transhumanist ladder (v0.3+) asks what you BECOME. But these two frameworks may be in tension. + +In CK3, you can't transcend your starting traits through play — you inherit and you pass on. The characters change but within mortal limits. The transhumanist ladder is a break from that model: going Higher or uploading to ANA is a qualitative change in what kind of being you are, not a quantitative improvement in your proficiency scores. + +Does the CK3 character creation budget system need to have room built in for the player to OUTGROW it? If going Higher means gaining capabilities that don't fit the proficiency framework — or means leaving the physical framework entirely — then the skill system can't treat the transhumanist ladder as just more skill points. It's a different category. + +The Jeroen-confirmed direction for v0.2 is: plant seeds, don't build it yet. My question is: what structural decision do you make in v0.2 character creation that ensures the ladder is still reachable in v0.3+? Does the skill system need an explicit ceiling that the ladder breaks through? Or does it need to be designed as inherently extensible — a budget that can be revised upward by the world, not just by the player? + +**Why this matters:** If character creation establishes a hard ceiling and the transhumanist ladder requires breaking it, that's a dramatic moment — you leave the person you designed behind. That's thematically rich. But it needs to be *designed* as a break, not accidentally foreclosed. The v0.2 architecture decision is whether going Higher is a modification of the existing character model or a replacement of it. + +--- + +## My Single Most Important Recommendation + +**Don't let the Phase 1 uncaring world become the game's default emotional register.** + +Kenshi is great because you're always fighting the indifference. If v0.2's Phase 1 is long enough and rich enough that the player emotionally settles into "this world runs without me," Phase 2's authored moral weight will feel like an intrusion rather than an escalation. The player will have accepted a world without consequence, and then the game will try to introduce one. + +The Phase 1 world needs to be built so that the player is *leaning into* the world, not resigned to it. Curiosity, not detachment. The right model is not "this world doesn't care about you" but "this world runs whether or not you engage — but if you engage, it responds." The responsiveness is latent in Phase 1, activated in Phase 2. The player should spend Phase 1 learning that the world is responsive, so that Phase 2's authored escalation feels like the natural consequence of having paid attention. + +This is a design note, not an architecture note. It changes what you author into Phase 1 — not the scripted arcs, but the ambient signals that the world registers your presence. Small things: an NPC who mentions you came in yesterday. A shop where prices shifted because you bought something. A door that's unlocked now because someone saw you helping. The world doesn't care about your moral weight yet, but it notices that you exist. That's the register Phase 1 needs to establish before Phase 2 can land. diff --git a/docs/workshops/wheres-the-fun/round4-interview.md b/docs/workshops/wheres-the-fun/round4-interview.md new file mode 100644 index 000000000..6fd2c42c0 --- /dev/null +++ b/docs/workshops/wheres-the-fun/round4-interview.md @@ -0,0 +1,106 @@ +# Round 4 Interview: Near-Miss Refinement +## Where's the Fun? Workshop | 2026-03-05 + +--- + +## Group 1: Proof-of-Life Scope + +### Q1 (Tyre): Minimum viable world — hand-built or procedural? + +**Jeroen:** "I like the idea in principle, but similar reasoning led us to not aim for more complex systems for v0.1. To be honest I am contemplating the proof of life should be that we can auto generate locations automatically (and therefore at reasonable scale) and then worry about the game. This includes better graphics, since the dots are really disconnecting." + +**Decision:** The proof-of-life is the GENERATOR + GRAPHICS, not a hand-built vertical slice. The v0.1 lesson: descoping led to the wrong game. v0.2 proves the foundation (auto-generated locations at scale + legible characters) first, then builds the game on top. + +### Q2 (Tyre): Character creation scope for v0.2? + +**Jeroen:** "Skills plus bookmark only for now. And I don't want religion to be a main driver in anything. I just mentioned that because of the importance in CK3." + +**Decision:** Skills + bookmark only. Religion is NOT a game system — it was a CK3 reference point, not a design requirement. Family/culture deferred. + +### Q3 (Tyre): Career models for v0.2? + +**Jeroen:** "I'd rather do tycoon in that case. I want nothing to do with the investigation stuff in this release. We'll revisit when the game stabilizes." + +**Decision:** Tycoon is the v0.2 bookmark. Zero investigation content. Clean break from the detective/smuggler framing. Tycoon naturally blends Active (manage business), WFH (remote investments via insert), and Gig (one-off deals). + +--- + +## Group 2: Skills & Voice + +### Q4 (Gestalt): Skills-verb coupling — priority, availability, or outcome? + +**Jeroen:** Mostly outcome (C). + +**Decision:** Everyone sees the same verbs. Skills determine how well you do. Bad at social? You can still talk, just badly. Some advanced verbs may still be gated by skill — spec needed for which ones. + +### Q5 (Mellanie): Voice attribution — character or job? + +**Jeroen:** "Job modifies culture. A character is their component experience. A job is an adder to this." + +**Decision:** Culture-driven voice, job modifies. INVERTED from Mellanie's option C. The character IS their background. Job adds a layer. Voice cards are authored at the culture level with job-specific modifiers. A Krenn tycoon sounds like a Krenn person who runs businesses, not a generic tycoon. + +--- + +## Group 3: Content Architecture + +### Q6 (Paula): Is Kael always Kael, or a generator-filled role? + +**Jeroen:** "This is exactly why I want auto generated first. The locations, archetypes and roles have become way too rigid in v0.1 out of simplification purposes. Kael should not exist. He should have been an auto generated NPC that fit the smuggler position because that particular location lent itself to smuggling because of its positioning. The Sims and Rimworld are perfectly capable of generating interesting characters that come alive and create attachments with (even without dialog in Sims and in the social panel or with mods in Rimworld with limited content). I realize the copy pool will get enormous, but you are generative AI that can deal with templating vectors like cultures, worlds, tones of voice, accent prompts to get this done in a varied way. I am okay with the 'vocabulary' being limited at first for NPCs. We will investigate templating for this, and maybe we can look at running a dressed down version of ollama with a relatively simple AI live in game for this. A problem for later." + +**Decision:** ALL NPCs are generated. No named characters. The generator produces NPCs that fit positions based on location characteristics. Content templating via generative AI (culture vectors, tone, accents). Possible in-game ollama for live NPC dialogue — deferred but the door is open. Limited vocabulary acceptable at first. + +### Q7 (Gore): Phase 1 emotional register — indifference or quiet responsiveness? + +**Jeroen:** "I reject Gore's premise. Authored content will be in before v1.0 so we are dealing with test users. Quietly responsive and less quietly responsive where we are dealing with primary social contacts (colleagues, neighbors over time). The world doesn't care, but it does notice and bits do start caring." + +**Decision:** Quietly responsive. The world doesn't care globally but notices locally. Primary social contacts (colleagues, neighbors) develop responsiveness over time. Gradient of caring based on social proximity. Gore's concern about indifference is addressed — the world will never be truly Kenshi-indifferent, even in early builds. + +### Q8 (Araminta): Character appearance customization depth? + +**Jeroen:** Full customization. Hair, clothing, colors. The character creation screen is part of identity investment. + +**Decision:** Full character customization despite top-down tile scale. Readability solved through outline/highlight, not by limiting customization. The creation screen is an emotional investment moment. + +--- + +## Group 4: World Identity + +### Q9 (Miri): Setting delivery — insert, physical world, or both? + +**Jeroen:** Both layered (C). + +**Decision:** World shows it through visuals and behavior; insert names and contextualizes. Araminta (visual) and Mellanie (insert copy) work in parallel. + +### Q10 (Miri): First Settled Reach moment? + +**Jeroen:** "Insert activation AND the apartment. The apartment will be auto generated (wealthy, poor, a rural start, will matter). The alarm clock, if that is legally allowed, I would want to start sounding like the *click* pa-pa pa-pa opening from Groundhog Day, cut short. Only at the beginning of the game to make the 'new day new start new chances' land with a wink." + +**Decision:** Two moments layered: (1) Waking up in YOUR auto-generated apartment (reflects your economic position). (2) Insert activation (neural implant powering on — intimate, personal, tech-specific). The alarm clock sound is a Groundhog Day homage — *click* pa-pa pa-pa, cut short, first game day only. "New day, new start, new chances" with a wink. + +### Q11 (Gestalt): Is Phase 1 dev-sequencing or player experience? + +**Jeroen:** "There will be authored content, but this is missing the point. The content is what the user goes and does. Every Rimworld game starts with a crash (well the first 3 DLC and unmodded does) then there is only agency and options beckoning. A job is a set of rails to take off from." + +**Decision:** The question is wrong-framed. Phase 1 isn't "empty world before content" — it's "world full of opportunity where the player's choices ARE the content." Rimworld model: one authored starting beat (the crash / the alarm clock), then agency and options. A job is rails to take off from, not a script to follow. The world provides opportunity and consequence; the player provides the story. + +--- + +## Summary: All Decisions From Round 4 + +| # | Decision | Impact | +|---|----------|--------| +| 1 | Proof-of-life = generator + graphics, not hand-built slice | Changes v0.2 milestone definition | +| 2 | Skills + bookmark only for creation | Defers family/culture/religion | +| 3 | Religion is NOT a game system | Removes from design scope | +| 4 | Tycoon is v0.2 bookmark, zero investigation | Clean break from v0.1 | +| 5 | Skills affect outcome (mostly C) | Simplest verb model, learnable | +| 6 | Voice: culture-driven, job modifies | Voice cards at culture level | +| 7 | All NPCs generated, no named characters | Eliminates hand-authored NPCs | +| 8 | Generative AI for NPC content templating | Opens AI-assisted content pipeline | +| 9 | Possible in-game ollama for live dialogue | Deferred but door open | +| 10 | Quietly responsive world, not indifferent | Gradient of caring by social proximity | +| 11 | Full character customization | Identity investment at creation | +| 12 | Setting: both layers (visual + insert) | Parallel production tracks | +| 13 | First moment: apartment + insert activation | Auto-generated apartment reflects status | +| 14 | Groundhog Day alarm clock homage | First day only, tonal wink | +| 15 | Player choices ARE the content (Rimworld model) | Job = rails to take off from | diff --git a/docs/workshops/wheres-the-fun/round4-mellanie.md b/docs/workshops/wheres-the-fun/round4-mellanie.md new file mode 100644 index 000000000..f12f020f8 --- /dev/null +++ b/docs/workshops/wheres-the-fun/round4-mellanie.md @@ -0,0 +1,118 @@ +# Round 4: Mellanie — Cross-Review and Refinement Questions + +**Workshop:** Where's the Fun? | **Round:** 4 (Synthesis) | **Agent:** Mellanie + +--- + +## Agreements + +**With Paula:** Phase Zero is right. The smuggler's arc can't begin until comfort exists, and comfort can't exist without life-texture monologue — lines that sound like muscle memory, not like someone entering a scene. Phase Zero content (ordinary shift, Kael warm and present, nothing wrong yet) is the first copy deliverable, not the last. + +**With Araminta:** Spatial anchoring is a prerequisite for content iteration. Q9 from the interview confirmed: the player knew the monologue was trying to communicate something, but couldn't connect it to source. Better lines in the same unanchored delivery context would have produced the same result. The source tile highlight comes before any monologue content push. + +**With Ozzie:** The "Character's Instinct" wow moment ("my character knows something I don't") is the best framing for what career-specific monologue actually does. Not "the character tells the player something" but "the character notices something the player hasn't consciously registered." That's 10 career-tagged anomaly lines per career — and it's concrete enough to write once voice is resolved. + +**With Gestalt:** VerbPriorityProfiles per career affect my trigger catalog. The anomaly triggers I write need to match what each career's verb profile actually surfaces. A detective who can't easily *Observe* at close range (Talk > Observe bug) would never trigger the observation lines I'm writing for them. Copy and verb system need to be coordinated. + +**With Nigel:** The cross-career comparison vision ("the tycoon's shell company was the smuggling ring the detective was investigating — neither knew") requires voice cards distinctive enough that the *same event* reads utterly differently through different career lenses. Five headlines × 3 career reactions isn't just a content deliverable — it's a test of whether the voice cards are working. + +--- + +## Tensions and Near-Miss Risks + +### The voice attribution problem + +The interview supplement changes the architecture of my domain and I don't know the full answer yet. + +The current voice cards (smuggler, detective) are **job-voices**. The CK3 model says character creation — skills, culture — is *more fundamental* than job. The job determines the toolbox (which inserts activate, which contacts exist, what information is available). It doesn't determine the personality. + +So: a Krenn-background social manipulator and a military-background enforcer both take the law enforcement job. Same toolbox. Different people. Do they share a voice card? + +If voice comes from character creation, the career voice card I've been building is the wrong level of abstraction. I'm writing voices for jobs when I should be writing voices for people. That's either a combinatorial explosion (culture × skill × career = how many voices?) or it means voice cards need to parametrize register rather than authoring full distinct pools. + +I don't know which Jeroen intends. This needs an answer before I produce more voice card work. + +### The three career models produce three different monologue cadences + +The interview supplement introduces three structural models: +1. **Active** — at the workplace, the work is the gameplay (bartender, patrol officer) +2. **WFH/Remote** — portable work done anywhere (hacking contracts, remote consulting) +3. **Gig/Freelance** — pick jobs from a board, episodic (smuggling runs, fixer contracts) + +These have fundamentally different monologue rhythms: +- Active career: continuous ambient commentary on the shift, the place, the people. Same location, day after day. The ordinary *is* the content. +- Remote career: focused inner monologue about the work itself — the target system, the client, the job in isolation from the physical world around them. +- Gig career: episodic decision-making. Pre-job (do I take this?), in-job (is this going sideways?), post-job (what just happened?). Job start and end are trigger events. + +The trigger catalog I proposed was written without this distinction. `job_event`, `consequence_visible`, `mission_outcome` map cleanly to Gig. They don't map as cleanly to Active or Remote. The Active career's monologue is mostly in the existing perception-event trigger space — it's ambient life, not episode beats. Remote is something else again. + +This isn't a blocker, but it means the trigger catalog has three variants, not one. I need to know which career model the v0.2 vertical slice uses before I finalize the catalog. + +### The world-first sequence creates a content gap + +The two-phase approach: Phase 1 is the uncaring world (no authored content). Phase 2 is when triangle templates, FRIEND arcs, contradiction arcs get injected. + +Monologue has to exist in both phases but doing different things: +- Phase 1 monologue: pure life-texture. *"Another shift. The recycled air still costs more on the dock floor."* No hooks, no investigation, no moral weight. +- Phase 2 monologue: anomaly recognition and consequence commentary. The lines that land after something authored enters the world. + +These are distinct content types requiring distinct trigger logic. Phase 1 lines fire continuously during ordinary play. Phase 2 lines fire on authored events. If we write them in the wrong order or mix them in the same pool, Phase 1 feels like a game waiting to reveal its hand rather than an actual life happening. + +The sequencing risk: if Phase 2 content is written first (because it's more interesting to write), we'll have a pool full of dramatic consequence lines with no Phase 1 life-texture underneath them. The moral arc needs Phase Zero. Phase Zero needs content. That content is Phase 1 monologue. + +--- + +## Questions for Jeroen + +### Q1: Does character creation (culture/skill) determine voice register, or does job determine it? + +This is the copy pipeline's most important unresolved question. + +The CK3 model says character creation is more fundamental than job. Job gives you the toolbox. But the toolbox's *voice* — how the character processes what they see, what they notice, what they care about — where does that come from? + +Options: +- **Job determines voice:** The detective sounds like a detective regardless of creation choices. Culture and skill affect what content fires, not how it sounds. Voice cards stay career-level. +- **Character creation determines voice:** A military-background character and a merchant-background character have different speech patterns, different emotional registers, even if they hold the same job. Voice cards need to be culture-level or skill-level. +- **Both layer:** The job provides the base register; character creation adds modifiers (word choice, emotional distance, what the character notices first). Parametric rather than fully authored. + +The answer determines whether I write 2 voice cards or 20 and whether the cards are fully authored or parametric templates. This is a production planning question that's currently unanswerable. + +--- + +### Q2: In the two-phase sequence — world runs first, content injects second — what's the first authored monologue content that enters, and when? + +Jeroen says: "Once the world runs and feels natural (but artificial) is when we start with authored quest and mission components." + +I read this as: Phase 1 runs without authored *mission* content. But monologue for Phase 1 (ordinary life, place texture, who you are before anything happens) might be some of the earliest authored content needed — not mission content, but voice content. + +Or does Phase 1 run completely silent, with the world establishing itself through visual and behavioral means alone, and monologue entering only when Phase 2 authored events begin? + +The distinction matters because: +- If Phase 1 needs monologue: I write life-texture lines first. Phase Zero content is the first deliverable. The voice establishes itself in silence before the drama arrives. +- If Phase 1 runs without monologue: I write anomaly/consequence lines first. The voice's first words are reactive, not ambient. + +The wrong choice produces either a silent Phase 1 that feels like a tech demo, or a Phase 1 monologue that feels out of place before any authored arcs are active. + +--- + +### Q3: Is the "gig/freelance" career model the right frame for the v0.2 vertical slice — and what does that mean for how missions are structured as "emergent things"? + +The three career models (active, remote, gig) have radically different content needs. Gig is the most mission-adjacent; Active is the most life-simulation-adjacent; Remote is somewhere in between. + +The interview supplement suggests the v0.2 slate includes "I have a plan, let's see if it flies" as much as formal missions. But from a copy standpoint, the difference between a gig worker's episodic job-board contract and a day of cooking at the bar while meeting people is not just tonal — it's a different trigger architecture, a different monologue rhythm, a different relationship to time. + +Is there a confirmed career model for the v0.2 vertical slice, or are we building content for all three simultaneously? If all three, the content workload triples and we need to prioritize deliberately. If one, I can focus the trigger catalog and voice cards on that model and mark the others as future work. + +--- + +## My Single Most Important Recommendation + +**Answer the voice attribution question before writing any new voice cards.** + +The v0.1 playtest failed in part because content was written for a world that didn't exist yet. The v0.2 risk is content written for a voice model that doesn't match the character creation system. + +If character creation (culture, skill, background) determines voice register, then career-level voice cards are the wrong primitive. We should be writing culture-level register parameters that job templates then specialize. If job determines voice, the current approach is right and I can produce voice cards as fast as the career list is confirmed. + +This is the copy equivalent of Tyre's "the mission system needs a design workshop before implementation." Write the design spec for the voice model before writing the lines. It takes a week at most. The alternative is a second pool of content written on a wrong assumption. + +The sequence: confirm voice attribution model → confirm career bookmark list for v0.2 → write voice cards → write Phase Zero life-texture content → wait for NPC legibility before building the full line pool. diff --git a/docs/workshops/wheres-the-fun/round4-miri.md b/docs/workshops/wheres-the-fun/round4-miri.md new file mode 100644 index 000000000..339ff9823 --- /dev/null +++ b/docs/workshops/wheres-the-fun/round4-miri.md @@ -0,0 +1,106 @@ +# Round 4 — Miri: Cross-Review and Refinement +## Where's the Fun? Workshop | 2026-03-05 + +**Domain:** Worldbuilding & Setting Design +**Reading:** All Round 3 proposals, Qatux Round 3 notes, interview supplement + +--- + +## Reactions to Other Agents' Proposals + +### Gestalt — VerbPriorityProfile per career + +Strong alignment. Setting note I want to add: verb priority isn't just a systems problem — it's a world-legibility problem. When a law enforcement officer approaches a citizen, the world should *respond to their authority* before they even use a verb. The Commission insignia on their insert, the NPC body-language shift from indifferent to cautious, the small behavioral tells that say "this person knows who you are." That's the setting layer underneath the verb priority spec. Gestalt's VerbPriorityProfile needs a worldbuilding input — what does each career's *relationship to the world's social structure* mean for how the world reads back at them? This is a joint design item. + +### Ozzie — 6 wow moments redesigned as emergent thresholds + +"First Day" (I belong somewhere) and "Ownership Moment" (that's MINE, someone is threatening it) are the two most worldbuilding-critical moments in Ozzie's revised list. Both require the setting to be legible before they can fire. "I belong somewhere" only works if "somewhere" reads as a specific place worth belonging to. The "Ownership Moment" only works if owning a bar stall in Sova Transit's workers' quarter means something different from owning a Commission contract. Strong alignment — but these moments are setting-dependent in a way the other four aren't. I'd add a dependency annotation: "Ownership Moment requires economic texture layer." + +### Paula — Phase Zero + +Paula's Phase Zero (warmth-establishing before the arc begins) now needs to map onto the Groundhog Day structure. The Groundhog Day model starts at Day 1 of onboarding — no pre-arc setup. This means Phase Zero IS the first several days of the career onboarding arc, not a separate designed beat. Paula's content (warmth with Kael, seeing Naia in his life, clean jobs) becomes the authored content injected into the first N days of the smuggler bookmark. This is a sequencing clarification, not a conflict. The content is right; it just needs to be expressed through the Groundhog Day cadence rather than as a prologue. + +### Gore — "weight of having lived" + +The supplement's Phase 1/Phase 2 sequencing is actually the strongest possible support for Gore's thesis. An uncaring world that runs without you — where the economy ticks and NPCs accumulate histories before you arrive — gives "the weight of having lived" a foundation. You're not a special protagonist who makes the world happen. You're a person in a world that was already happening. That indifference is the precondition for Gore's consequence thesis to feel earned rather than scripted. Strongly aligned, and I think Gore should note this explicitly in their Round 4 response. + +### Nigel — career divergence as replayability spine + +The supplement's character creation model (CK3-style skill budget, not archetype class selection) adds a dimension Nigel's proposal didn't anticipate. Two players both playing law enforcement but with different skill builds — one social-manipulation-heavy, one combat-proficiency-heavy — will have different games within the same career. Replayability isn't just career divergence; it's build divergence within careers, amplified by world-state randomness. Nigel's framework holds, but the combinatorial space is larger than described. + +### Tyre — architecture + +The Phase 1 "uncaring world" model is what the Rust simulation server was built to be. Strong alignment throughout. One question I need Tyre to answer (see below): does the generator's zone type taxonomy carry enough information to produce place identity, or does it need a worldbuilding spec layer that feeds zone rules? + +### Araminta — visual hierarchy + +The Groundhog Day opening changes Araminta's priority ordering slightly. The first visual frame — what the player sees when they "wake up" — is now critical. Before any calendar ping resolves, before any NPC interaction, the physical world in the camera frame needs to communicate "this is Sova Transit, and you live here." Araminta's functional cluster palette system needs to produce that legibility not for an oriented player already exploring, but for a just-woken-up player with zero context. The insert AR overlay is also part of this first frame — it's what the player sees *as their character* sees it. The career insert visual grammar document needs to spec what the very first insert activation looks like. + +### Mellanie — monologue as interior commentary + +The Groundhog Day model has an immediate implication for monologue: the character's first line should fire on waking up, not on perceiving something. "Another day at Sova Transit" or whatever the voice card produces for that career + situation. The very first line establishes that this character has a relationship to this place — before the player has done anything, the character already has a history here. Mellanie's life-event trigger expansion needs `day_start` as a trigger type. Strong alignment with the worldbuilding goal of making the setting feel inhabited from second one. + +--- + +## Setting-Specific Conflicts I'm Watching + +### The generator and world identity — not yet addressed + +The supplement names "the Generator Architecture workshop" and "Cities Skylines top-down pipeline model with 14 confirmed D-records" as established. But no Round 3 agent addressed the worldbuilding layer of the generator. Cities Skylines produces legible city space because its zone rules encode what residential/commercial/industrial look like. Our generator needs equivalent rules for what a logistics hub in a working-class transit district looks like vs a Commission administrative zone vs a bar quarter. + +This is a gap: the Phase 1 "uncaring world" cannot produce Sova Transit specifically without worldbuilding rules feeding the generator. The generator knows zone types. It needs to know what zone types *mean* in the Settled Reach's social vocabulary. + +**This is the most important unresolved worldbuilding item from Round 3 + the supplement.** + +### Setting work dependency chain (Qatux tension #5) + +Qatux flagged this and it's real. Miri writes the Sova Transit place identity document → Araminta produces the tile palettes and NPC sprites → Tyre exposes the zone-type data in the snapshot → generator feeds zone-type rules. These are sequential dependencies. The Phase 1 uncaring world can't produce legible Sova Transit until all four are in place. This chain needs to be made explicit in the sprint planning — it's not parallel work. + +--- + +## Three Questions for Jeroen + +### Question 1: What does the generator need to know to produce "the Settled Reach" rather than generic sci-fi urban space? + +The supplement establishes Phase 1 as the generator running independently: Geography → infrastructure → zones → population → routines. In Cities Skylines, zones encode visual and functional identity — an industrial zone looks and behaves differently from a residential one before any author touches it. + +For the Settled Reach's generator to produce Sova Transit as a specific kind of place — working-class, transit-adjacent, Commission-surveilled, economically anxious — what rules does the generator need to hold? Is the zone type taxonomy (logistics, residential, commercial, administrative) sufficient to produce place character? Or does the generator need a richer worldbuilding input — a district identity spec that says "logistics zones here have visible span gate infrastructure, Commission checkpoints, and cargo-handling NPC routines"? + +**Why this matters:** If the generator can produce recognizable Sova Transit from zone types alone, the worldbuilding layer is primarily an authoring problem (content and copy). If the generator needs richer zone rules to produce the right feel, that's a design spec I need to write before the generator can run Phase 1 correctly. I need to know which side of that line we're on. + +### Question 2: In the Groundhog Day opening, what is the player's first visual impression — and what does it need to communicate? + +The alarm clock fires. The player's camera comes up. Before the calendar ping resolves and before any NPC is visible, what do they see? Their apartment, a view of the district, the morning routine playing out in the world? + +And from a setting perspective: what does that first visual frame need to communicate about where they are? Is the setting communicated through the insert AR overlay (the character's subjective digital layer over the world, which could tell you your address, what you owe, what the day holds)? Or through the physical world in the camera (the industrial aesthetic of Sova Transit, the span gate visible in the distance, the morning shift arriving at the logistics hub)? Or is the expectation that both work simultaneously? + +**Why this matters:** These two channels (insert AR vs physical world) are different worldbuilding and production problems. The insert AR layer is authored copy and UI — a few lines of text about your life. The physical world is visual design, tile work, ambient population. If the first 30 seconds depends primarily on the insert AR layer, that's a faster path to setting legibility with lower art dependency. If it depends on the physical world reading clearly, that's Araminta's work and mine in parallel before the Groundhog Day opening can land. + +### Question 3: What makes the first session feel like the Settled Reach specifically, not a generic life-sim skin? + +The confirmed reference games — Sims, CK3, Rimworld, Dwarf Fortress, Kenshi — are the right structural inputs, and the surface must remain wholly original. But the combination of day-cycle + job + relationships + job board is a genre pattern a player will recognize. The question isn't IP risk from specific franchises; it's genre-blur risk. A player picking up the game cold might feel "oh, this is basically [life sim genre]" before they feel "this is the Settled Reach." + +What is the element in the first session that is *only possible in this world*? My candidates: +- The **insert activation** — the neural technology that surfaces different information depending on your career, that feels like a part of your body and costs money to upgrade +- The **span gate visible from your window** — the reminder that you're working-class at the bottom of the access hierarchy, that the gate exists and you can't afford it +- The **Commission presence as ambient fact** — not a quest giver, just visible authority that the world organizes itself around + +Which of these (or something else) should be designed as the *first distinctively Settled Reach moment* — the thing that signals this is not Stardew Valley with a sci-fi skin? + +**Why this matters:** IP originality is my standing mandate. We're safe from direct franchise copying, but genre-blur is a different risk. The answer to this question should become a design principle for all bookmark onboarding arcs: "by the end of Day 1, the player has experienced X, which only happens in the Settled Reach." + +--- + +## My Single Most Important Recommendation + +**Write the zone identity spec before the generator runs.** + +The Phase 1 "uncaring world" is the foundation for everything else. But a generator that produces zone types without knowing what those zones *mean in the Settled Reach* will produce generic space, not Sova Transit. The worldbuilding layer that feeds the generator — what does a logistics zone look, sound, and behave like in this specific world? — is upstream of Araminta's tiles, upstream of Tyre's snapshot data, upstream of Mellanie's first monologue line. + +This spec is one document, probably one sprint, and it unlocks the entire Phase 1 foundation. It should be the first worldbuilding deliverable in the v0.2 roadmap, not the Sova Transit place identity document I proposed in Round 3. Those are the same thing — I'm just naming the dependency more precisely now. + +**The spec answers: what rules does the Settled Reach's generator need to produce a world that reads as this world?** + +--- + +*Miri — Round 4 complete.* diff --git a/docs/workshops/wheres-the-fun/round4-nigel.md b/docs/workshops/wheres-the-fun/round4-nigel.md new file mode 100644 index 000000000..bb511aba7 --- /dev/null +++ b/docs/workshops/wheres-the-fun/round4-nigel.md @@ -0,0 +1,117 @@ +# Round 4 — Nigel: Cross-Review and Refinement + +**Workshop:** Where's the Fun? v0.1 Playtest Reckoning +**Agent:** Nigel (Sandbox & Replayability) +**Round:** 4 — Synthesis + +--- + +## Reactions to Other Agents' Proposals + +### Gestalt — VerbPriorityProfile per career + +Strong agreement. This is replayability infrastructure that doesn't look like it, but is. If a law enforcement detective and a smuggler encounter the same NPC and the verb priority system presents them with genuinely different default actions, those playthroughs feel like different games from the inside, not just different information. The verb layer is where career-as-lens becomes tactile. Keep this in v0.2 scope. + +### Ozzie — The Ownership Moment + +The best new wow moment. "That's MINE. Someone is threatening it." An owned asset with a threat state is a consequence cascade machine — it creates permanent stakes, it makes prior choices feel weighty, and it's the most natural replay-desire generator: what would I have owned if I'd played differently? The Ownership Moment should be prioritized alongside the Consequence moment, not after it. + +### Gore — "Consequence" over "complicity" + +This reframe directly reinforces the replayability architecture. If the organizing concept is consequence — choices have mass, they accumulate, they shape you — then consequence cascades are thematically coherent, not just mechanically useful. "The game is about the weight of having lived" is the right sentence to put on the box. And it's exactly what the comparison test produces: two players with different weights, shaped by different choices, comparing what they each ended up carrying. + +### Tyre — Mission system as highest-risk new system + +Correct. The consequence cascade engine I described in Round 3 is only as good as the mission system feeding it. If missions are binary (success/fail → clean consequence mapping), the emergent life stories are too predictable. If missions have genuine outcome spectrums with downstream effects that aren't fully legible until later, they produce real surprises. Tyre's flag about needing a design workshop before implementation is the most important process recommendation in the round. I second it emphatically. + +### Mellanie — Hold new monologue until NPC legibility is solved + +The right call. And it has a replayability dimension that nobody else mentioned: monologue that references specific people only lands if the player knows who those people are. Career-specific monologue reacting to life events (relationship shifts, financial consequences, mission outcomes) is the version that creates replay desire — hearing how your character processes the specific life they've built. But that requires both legible NPCs and a life worth processing. The precondition is legibility, not content volume. + +### Miri — Bookmark onboarding as worldbuilding delivery + +This is the thing that makes career divergence legible to the player. If each career path enters the same Sova Transit from a different social position and each reveals a different face of the world, the replayability isn't just structural — it's experiential. The dock worker doesn't know the bar exists. The bar worker doesn't know the dock routing manifest exists. The Commission detective walks through places the others inhabit and reads them as evidence. Same world, different social entry points, completely different place. + +### Paula — Phase Zero before the moral arc + +Essential. The emotional weight of consequence cascades requires an earned foundation. You can't feel the weight of what you did if you were never comfortable enough to have something to lose. Phase Zero isn't just narrative setup — it's the replayability precondition. The second run's asymmetric lens only works if the first run built real attachment. + +--- + +## The New Complication: Character Creation Precedes Career + +The interview supplement changes my Round 3 framing in an important way. + +I argued that career bookmarks are the primary replayability source. Jeroen's direction says character creation — skills, culture, family, budget — is MORE fundamental than career choice. The job/archetype determines the toolbox. The character determines who's using it. + +This is actually BETTER for replayability than career divergence alone. But it raises a design question I haven't seen anyone address: **does character build interact with career to produce genuine game-state divergence, or does career trump build in terms of what the world offers?** + +If a high-social law enforcement character and a high-hacking law enforcement character have genuinely different games — the social character builds informant networks, the hacker extracts evidence from insert comms, they're investigating the same ring through completely different angles — then the replayability state space isn't n careers, it's n careers × m meaningful build combinations. That's enormous. And it nails replay variety without us engineering it. + +But if career is primary and character build is flavor within a career (the hacker detective still follows the same case structure, just opens some doors differently), then build variety doesn't add structural replayability. It adds tactical variety. Different, but less. + +This distinction matters architecturally. Career divergence = different games. Build variation within a career = different playstyles of the same game. Both good; not the same thing. + +--- + +## The Clean Start Risk + +The Groundhog Day framing establishes a "clean start" mode — no bookmark, no appointment, figure it out from the job board. Explicitly described as "experienced player warning." + +From a replayability standpoint, clean start is theoretically the highest-replayability mode: no pre-authored onboarding constraining the start, pure world-state gravity generating pull. But it's also the mode with the most acute version of the v0.1 problem — no signal, no pull, nothing indicating where to go or what matters. + +The question is whether Phase 1 of the two-phase world generation (the uncaring world running naturally) generates enough ambient pull for a clean-start player. The economy ticks. NPCs have routines. Job boards post. But does the player have enough world-reading tools to discover a foothold without authored onboarding content? + +I don't think we know the answer to this yet, and we shouldn't find out in the v0.2 vertical slice. Clean start should ship after one bookmarked career is fully working and the world is legible enough to be read without structured introduction. Otherwise we're recreating the v0.1 testing wall in a different framing. + +--- + +## The Near-Miss Risk for Replay + +The two-phase world generation is the right architecture. But it contains a replay risk I want to flag. + +Phase 2 injects authored ingredients (friend arcs, contradiction arcs, triangle templates) into the running world. The storyteller determines when and how much pressure, not what happens. This is correct. + +But if the authored ingredients are placed into the world independent of career path — random distribution regardless of who the player is — then two law enforcement runs will encounter the same authored skeleton (the ring, the corruption, the manifest discrepancy) through the same career tools. The variety is world-state variation only (is the ring newly formed or entrenched?), not structural story variety. + +The replayability only explodes if the authored content distribution is career-aware. A law enforcement run should surface authored ingredients that are legible to a law enforcement lens — and leave invisible the authored threads that would only be visible to a smuggler. Two law enforcement runs hit the same pool of potentially visible content; two different-career runs hit structurally different content pools. + +This is the difference between: +- **World state variation:** Same authored skeleton, different timing and pressure (good for within-career replay) +- **Career-aware content distribution:** Different careers activate different authored ingredients from the same world (the cross-career comparison test that produces "we were in the same simulation and saw completely different things") + +Both are needed. Neither alone is sufficient. And the career-aware content distribution is harder to design, which is why I'm flagging it now before implementation. + +--- + +## Questions for Jeroen + +### Q1: Does character build create structural divergence within a career, or tactical variation? + +You said skills (shooting, hacking, social manipulation, mechanical repair) are proficiencies, and the job/archetype determines the toolbox. What I need to understand: if two law enforcement characters build differently — one maxing social manipulation, one maxing hacking — do they experience structurally different games within the same career? (Different information surfaces, different authored ingredients becoming visible, different relationship networks becoming accessible?) Or do they play the same law enforcement story through different tactical approaches? + +This matters because: if high-social vs. high-hacking within law enforcement produces genuinely different games, the replayability state space is career × build combinations. If it's tactical variation within the same career structure, the replayability comes primarily from career divergence, and character build is flavor. + +Both designs are viable. But they imply completely different scopes for the character creation system and the world state architecture. I need to know which one we're building toward. + +### Q2: In the two-phase generation, does the storyteller place authored content career-aware — or is placement distribution-first, with career determining the player's lens on what's there? + +The Rimworld storyteller determines when and how much pressure, not what happens. That's correct for dramatic pacing. But for the cross-career comparison test to work — the tycoon discovering they financed the ring the law enforcement player investigated — the authored content needs to be placed such that it's partially visible from one career lens and fully visible from another. + +Is the vision: all authored ingredients exist in the world at all times, career determines which you can perceive? Or: the storyteller seeds different authored ingredients depending on career path, so a tycoon run genuinely has different active storylines than a law enforcement run in the same world state? + +The first is simpler and more emergent. The second is more controlled and more reliable for producing the cross-career story payoff. Which direction are we building? + +### Q3: The clean start mode is described as "experienced player warning." What's the design intention for when a player uses clean start — is it sandbox mode for players who know the world and want unstructured play, or is it a second-run mode for players who want to see what the world looks like without an authored entry point? + +This distinction changes how I design for it. If clean start is sandbox mode, it needs sufficient world legibility that a player who knows the systems can find their own foothold. If it's second-run mode, it needs the world to be generating enough ambient pressure that a player who did a bookmarked first run can find threads to pull from prior experience. These are the same system doing different jobs — I want to know which one we're optimizing for. + +--- + +## My Single Most Important Recommendation + +**Design the career bookmark list before anything else.** Qatux correctly identified this as blocking voice cards, visual insert designs, VerbPriorityProfiles, world state variables, and onboarding arcs. I'd add: it also blocks my ability to specify what world state variables matter, because which variables are relevant depends entirely on which careers are in the game. + +The world state for a law enforcement career (ring entrenchment level, evidence trail state, corruption network depth) is completely different from the world state for a tycoon career (market phase, faction economic power, available assets). The storyteller can't be designed until we know what careers exist. The consequence engine can't be scoped until we know what kinds of missions each career generates. + +The career list confirmation is the first design decision that unblocks everything downstream. It should happen in the next sprint before any implementation work on bookmarks, missions, or onboarding begins. diff --git a/docs/workshops/wheres-the-fun/round4-ozzie.md b/docs/workshops/wheres-the-fun/round4-ozzie.md new file mode 100644 index 000000000..605920149 --- /dev/null +++ b/docs/workshops/wheres-the-fun/round4-ozzie.md @@ -0,0 +1,121 @@ +# Round 4 — Ozzie: Cross-Review and Refinement Questions +## Where's the Fun? Workshop | 2026-03-05 + +**Agent:** OZZIE (Player Experience & Wow Factor) +**Round:** 4 — Synthesis + +--- + +## Reactions to Round 3 Proposals + +### The Convergence Is Real — and Slightly Too Neat + +All 9 agents agree. That's unusual. And it's mostly right. But I want to name one place where the agreement might be covering a real tension before we proceed. + +Everyone said "NPC legibility first." Everyone said "career bookmarks replace seed variants." Everyone demoted monologue. Everyone pointed at the diegetic tool suite. The consensus is genuine. But the *sequencing* implications of that consensus are NOT resolved — and three agents (Paula, me, Miri) each described a different "first thing" for a player's first 30 minutes: + +- Paula: establish warmth with Kael before any arc fires +- Ozzie: supervisor hands you a tool before anything emotional lands +- Miri: the world teaches itself through inhabiting it + +These aren't in conflict. But they imply a BEAT SHEET that nobody wrote. Qatux flagged it as a blocker. I'm flagging it as the most important unresolved player-experience question in the whole workshop. + +--- + +### Agent Reactions + +**GESTALT — VerbPriorityProfile per career:** LOVE IT. My "Character's Instinct" moment (monologue flags something the player missed) fires through career-specific observation verbs. The detective's close-range verb priority was broken. Per-career profiles fix this architecturally. Strong agreement. + +**PAULA — Phase Zero before Phase 1:** RIGHT. And here's what I want to add: Paula's "Phase Zero" and my "First Day" are the SAME beat described from different angles. Paula is thinking about relationship arc structure. I'm thinking about identity and belonging. The player needs both simultaneously — a role AND a person to be that role around. These should be designed as one unified opening arc, not two separate phases with different authors. + +**GORE — "Consequence" as the organizing concept:** Beautiful word choice. This directly names my Moment 3 ("That was ME") in its full meaning. I'm adopting it. The full wow moment stack can be described as: "You didn't solve a puzzle. You LIVED here. And living has consequence." That's the vision statement for this game. + +**NIGEL — Three replayability layers:** The career divergence → consequence cascades → world state randomness stack maps almost exactly onto my new wow moments: Asymmetric Lens, The Consequence, and the general texture. Nigel and I are describing the same emotional architecture from different angles (his from replayability, mine from first-run experience). Strong alignment. + +**MIRI — Setting legibility as designed deliverable:** My NPC legibility demand CANNOT happen without Miri's place identity work. These are coupled dependencies, not independent tracks. Araminta can design archetype silhouettes. But the silhouettes need to communicate roles — and role is a worldbuilding artifact, not a visual design artifact. One sprint: Miri writes the behavioral vocabulary, Araminta renders it. + +**TYRE — "Easy. The data exists. We just need to send it.":** This single sentence is the most important sentence in all of Round 3. My demand #1 (NPC legibility) is achievable in 1-2 weeks server-side. The bottleneck is visual design and authored content, not architecture. That changes the critical path. We're not waiting on a new system. We're waiting on decisions. + +**ARAMINTA — Career insert visual identity:** The moment your law enforcement insert activates with Commission blue and a case-file aesthetic — THAT IS A WOW MOMENT. The visual First Day. I should have listed this as its own beat. The insert activation sequence isn't just tutorial — it's the moment you see your career through a designed lens for the first time. Strong agreement and a gap in my own Round 3 I want to close. + +**MELLANIE — "Voice that voices rather than informs":** This distinction is CRITICAL for my "Character's Instinct" moment. The monologue has to react to what the player already perceives — not deliver data the player doesn't have yet. If the monologue lines are written as information delivery, the wow moment fails. If they're written as a second opinion on something you already saw, the wow moment fires. Mellanie has the most direct leverage on whether Moment 2 works. + +--- + +## The One Gap Nobody Addressed + +The supplement reveals that character creation is MORE fundamental than job selection. CK3-style: skills, culture, family, budget. The bookmark (job + location + relationships + tools) comes AFTER you've built the person. + +Nobody designed character creation as an EMOTIONAL EXPERIENCE. Everyone treated it as setup. That's wrong. + +In CK3, the character creation screen is where you first fall in love with your ruler. You give them a face, a background, a trait. By the time you hit play, you already care what happens to them. THAT'S how you solve the dots-aren't-people problem from the supply side — the player MADE this person. + +Character creation isn't setup. It's Wow Moment Zero. + +--- + +## Questions for Jeroen + +### Question 1: Is character creation a wow moment, or just setup? + +**The question:** In the Groundhog Day structure — alarm clock, calendar ping, appointment — when does the player first feel like they're playing A PERSON rather than controlling A CURSOR? Is it when they finish character creation and look at who they made? Or is it later, in the world, once the world responds to them? + +CK3's character creation makes the player invest before they play. The Sims' character creation does the same. If character creation is truly MORE fundamental than bookmark choice, it needs to be designed as an emotional experience — with a payoff moment at the end where the player looks at what they made and thinks "yes, that's who I am." Or is character creation here intentionally neutral, a canvas the player fills in through play? + +**Why it matters:** The answer determines where Wow Moment Zero lives. If character creation is an emotional beat, it's the first thing I need to design. If it's neutral setup, then First Day (alarm clock, supervisor, tool activation) carries the full weight of first attachment. + +--- + +### Question 2: What is the wow moment for the clean-start player? + +**The question:** The Groundhog Day structure gives bookmarked players a first appointment — supervisor, tool activation, onboarding. But the clean start (no appointment, experienced player warning) drops a player into the world with nothing but a job board. That player has no "First Day." No one handing them tools. No identity anchored to a place. + +The supplement says clean start is for experienced players. But will first-run players pick it anyway because it sounds like freedom? And even for experienced players: what is the WOW MOMENT in a clean start? Where does the "I belong somewhere" feeling come from when you chose to belong nowhere? + +Is the clean start supposed to have its own emotional arc — the found-family version, where belonging is earned not given? Or is it genuinely a power-user mode that we accept will have a rougher floor? + +**Why it matters:** If the clean start is an accessibility trap for first-run players, we need a warning strong enough to redirect them without feeling like a rebuke. If it's supposed to have its own wow moments, I need to design different beats for it than the bookmark path. + +--- + +### Question 3: How do I feel the Consequence in a Groundhog Day structure? + +**The question:** My "Consequence" moment (Moment 3 in my Round 3) is the first time a choice from a previous day shows up as a consequence in the current day. You hired someone on Day 3. Day 7, their mistake is YOUR problem. That's when the world proves it remembers. + +But the Groundhog Day structure starts each day with an alarm clock and a fresh calendar. How does the player FEEL the weight of previous days in a structure that foregrounds the new day? What is the in-world mechanism for "yesterday happened"? + +Is it the journal (Tyre's thread tracker)? An NPC who remembers your name and their expression has changed? A notification in the insert that references the earlier choice explicitly? Or does the Groundhog Day aesthetic specifically NOT include that weight — and consequence accumulates on a longer timescale than I'm imagining? + +**Why it matters:** The consequence moment is the pivot where the player stops exploring and starts LIVING. If the Groundhog Day structure mutes that pivot by emphasizing daily freshness over accumulated weight, I need to know now — before anyone writes onboarding content around that beat. + +--- + +## My Single Most Important Recommendation + +**Design the first 30 minutes as a unified arc, not a domain portfolio.** + +Right now, the first 30 minutes is being designed by four different agents from four different angles: +- Paula: Phase Zero (relationship establishment) +- Ozzie: First Day (identity and belonging) +- Miri: worldbuilding delivery (place legibility through inhabiting) +- Gestalt: diegetic insert onboarding (tool teaching through use) + +These are all RIGHT. But they're being designed independently, which means they'll arrive as four separate deliverables that don't know about each other. + +The first 30 minutes needs ONE beat sheet that all four domains design together, simultaneously. That beat sheet is probably: + +1. **Character creation** (Wow Moment Zero — you made a person) +2. **Day 1 alarm clock** (world anchors you in time and space) +3. **Appointment arrival** (supervisor, community, first look at your tools) +4. **Insert activation** (your career lens on the world, visual + functional) +5. **First work moment** (active: you're doing the job; the world responds to you doing it) +6. **First anomaly or texture beat** (either Phase Zero warmth OR a Sova Transit sensory moment OR your insert flagging something) +7. **First consequence seed** (a choice you make in these 30 minutes that will echo later) + +That's the beat sheet. Four domains. One arc. One design session. The wow moments can't be emergent until the floor is authored. + +THAT'S THE GAME. Let's build the floor before we celebrate the ceiling. + +--- + +*Ozzie — Round 4 complete.* diff --git a/docs/workshops/wheres-the-fun/round4-paula.md b/docs/workshops/wheres-the-fun/round4-paula.md new file mode 100644 index 000000000..c08708021 --- /dev/null +++ b/docs/workshops/wheres-the-fun/round4-paula.md @@ -0,0 +1,108 @@ +# Round 4: Paula — Cross-Review and Refinement Questions +## Where's the Fun? Workshop | 2026-03-05 + +--- + +## Reactions to Round 3 Proposals + +### Where I agree strongly + +**Ozzie's wow moment redesign** maps almost perfectly onto my Phase Zero proposal. "First Day" ("I belong somewhere") is exactly the emotional state Phase Zero is designed to deliver. "The Consequence" ("That was ME") is the Phase 1→2 transition given a design target. "The Asymmetric Lens" is the dual-lens property I said to stop engineering and let emerge. Ozzie's framework and my arc proposal are describing the same emotional beats from different angles — Ozzie from experience design, me from narrative structure. They should be designed together. + +The one sequencing note: Ozzie's priority order (NPC legibility → First Day → Consequence → Career monologue → Enemy → Lens → Ownership) is correct. Phase Zero cannot be authored until NPC legibility is solved. You can't write warmth about a person who doesn't visually register as a person. + +**Gestalt's VerbPriorityProfile reframe** matters for the moral arc. The current verb spec has Talk > Observe at close range — which is precisely wrong for Phase Zero. The smuggler's warmth with Kael should accumulate through repeated presence and observation before it deepens through dialogue. Phase Zero is watch-first, talk-later. Job-aware verb priority that demotes Talk for the smuggler (and amplifies Observe) directly supports the arc's pacing. + +**Mellanie's monologue redefinition** ("interior commentary on a legible world") is the correct framing and I want to reinforce it. The expansion of trigger catalog to include `relationship_shift` and `consequence_visible` is what makes Phase Zero content possible — you can't write "Kael makes a joke about the customs officer, warm, specific" as a perception trigger. It needs a `first_meeting` or `established_relationship` trigger type. This is a content architecture dependency I should flag to Mellanie directly. + +**Tyre's roadmap** is reassuring: Phase 1 (NPC legibility data) → Phase 2 (thread tracker, journal, bookmark definitions) → Phase 3 (onboarding sequences). The dependency order is correct for my domain. Phase Zero lives in Phase 3, which means it can't be fully authored until Phases 1 and 2 are done. I accept that sequencing. + +### Where I want to complicate + +**Gore's "consequence" vs "complicity" reframe** is useful but I don't think it's a full replacement. Let me complicate this. + +Gore's right that "consequence" is more accessible as a universal theme — choices have weight, accumulation shapes you, the game is about "the weight of having lived." That's true for tycoons, law enforcement, fixers, everyone. + +But "complicity" names something specific that "consequence" doesn't reach: *the realization that you were part of something before you understood what it was*. The smuggler's Phase 1 rationalization ("it's just logistics, nobody's getting hurt") isn't a wrong prediction — it's a way of not-knowing that the player participates in. The arc doesn't reveal consequences the player didn't cause. It reveals *what they were already doing*. + +My proposal: consequence as the universal theme. Complicity as the dominant register for careers built on entanglement in systems (smuggler, fixer, low-level law enforcement working dirty precincts). Aspiration/ambition as the dominant register for careers built on building things (tycoon, entrepreneur). The emotional arc varies by career; the underlying structure (Phase Zero → earned comfort → crack → reckoning → compromise) generalizes. + +Gore and I are not in conflict on the mechanism. We may be in slight disagreement about terminology. Resolve in the game's tone, not in this document. + +**Miri's two bookmark arcs (law enforcement + one other)** — I want to advocate for smuggler as the second. The most authored arc infrastructure exists for the smuggler path (Kael, Naia, the 4-phase moral arc, THE FRIEND pattern, Phase Zero content). Law enforcement makes sense as the "most legible" career per Ozzie (authority is simple to communicate visually, the job is self-explanatory). Smuggler is the one where we have the most designed emotional depth to deliver. Both together give the v0.2 vertical slice maximum contrast — the same world, opposite knowledge states, opposite moral positions. + +The open question (Q-WTF-008: law enforcement or smuggler for v0.2 vertical slice) might have the answer "both, but designed in the right order." Law enforcement first for legibility scaffolding, smuggler second for emotional depth. Or do both in parallel. This needs Jeroen's call. + +### The dependency I want to flag for the team + +Mellanie says "hold new monologue line-pool content until NPC legibility is solved." I agree as a principle, but Phase Zero warmth lines and NPC visual design need to be *co-specified*, not sequenced with visual first and content second. + +Why: Kael's warmth in the monologue depends on what Kael looks like and how he moves. If the visual design gives him a specific behavioral tell (a way of standing, an idle animation that communicates ease), the warmth lines should reference that tell. If the visual design reveals him through observation (you can tell he's a dock worker before he speaks), the monologue should echo what the player visually read. Content and visual design that don't know about each other produce a character who sounds different from how they look. + +This isn't a blocker on NPC legibility. It's an argument for a design handoff between Araminta/Miri and Mellanie/me before Phase Zero content is written. + +--- + +## Questions for Jeroen + +### Question 1: Does CK3 character creation change the *voice* of the moral arc, or just its pace? + +The interview supplement's CK3 reframe is exciting and changes the content architecture question significantly. + +In the current moral arc spec, the smuggler's Phase 1 voice is a single register: pragmatic rationalization, operational confidence, warmth toward Kael. It assumes one type of person doing the smuggling. + +But CK3 character creation gives two players the same "smuggler bookmark" while producing meaningfully different characters — different backgrounds, cultures, family histories, skill distributions. A player who came up through Burnelli freight logistics might rationalize Phase 1 as "this is just how logistics works, every freight network has gray zones." A player from a working-class Krenn background might frame it as "the Commission taxes everything that moves and calls it law — we just don't file." + +Both are in Phase 1. Both have the same arc. They sound completely different. + +**The question:** Does character creation affect the *voice register* of moral arc phases — specifically Phase 1 rationalization and Phase 4 compromise — or does voice differentiation happen only through the base voice card, with moral arc content staying constant? + +This matters because: if moral arc voice should vary by character background, then the content pool needs a second tag dimension (character background/culture) alongside the existing phase and trigger tags. That's a significant authoring scope expansion. If moral arc voice stays constant regardless of background (the arc is the arc, the voice card handles the rest), the existing content architecture is sufficient. + +### Question 2: In the two-phase world model, is Kael always Kael — or is he a role the generator fills? + +The supplement establishes a two-phase approach: world first (uncaring, emergent), then authored content injected. The generator places authored ingredients (triangle templates, FRIEND arcs) into the living world. + +The current moral arc spec names specific NPCs: Kael Davan, Naia Tamm, Maret, Pell, Voss. The authored monologue lines reference Kael by name, describe his specific behavioral tells, his relationship to Naia, his dock position. This is deeply specific authored content. + +In the generator model, when a smuggler bookmark is active, does the generator: + +**Option A — Named placement:** Always assign Kael Davan to the player's dock contact role. Kael is authored, hand-placed, and will always be the same person. The generator ensures he's present near the player's starting situation. His arc plays out as designed. + +**Option B — Role template:** Assign *whoever is in the appropriate social position* (dock contact, willing ring participant, person with a vulnerable partner) to the FRIEND-figure role. "Kael" becomes a template the generator fills with a named, generated NPC. The moral arc FactIds become generic (`smuggler.FRIEND.stress_escalation` rather than `smuggler.witnesses.kael_stress`). Authored lines become templates. + +Option A preserves the emotional specificity of the current arc but may feel artificial in a world that's supposed to be procedurally alive. Option B is more generative but requires a template-based content architecture rather than the authored-specific one we've been building. + +**The question:** Which direction is intended? This has cascading implications for content authorship — Option A is what's designed now; Option B requires a content architecture rewrite. + +### Question 3: Is Phase Zero a designed arc across days, or an emergent state the system recognizes? + +The Groundhog Day opening (alarm → calendar ping → onboarding location) gives Phase Zero a concrete frame. Day 1: meet Kael at the dock. Days 2-N: run jobs, build warmth, see Naia's presence in Kael's life. Eventually Phase Zero is complete and the Phase 1→2 gates become active. + +But "eventually" is doing a lot of work. How does the system know Phase Zero is complete? + +**Option A — Designed arc (scripted beats):** Phase Zero is a mini-arc designed across the first N sessions. Specific events are scripted: Day 1 you meet Kael, Day 3 you see Naia at the bar, Day 5 the first job goes slightly rough. The `smuggler.has_established_ring_relationships` prerequisite is set to true when the designed beats have fired. + +**Option B — Emergent threshold:** Phase Zero is a relationship metric. Repeated interactions with Kael accumulate a warmth value. When Kael's relationship score crosses a threshold, `has_established_ring_relationships` becomes true and the Doubt gates become active. No scripted beats — just the player's organic interaction rate driving the prerequisite. + +**Option C — Hybrid:** Day 1 is a scripted beat (the Groundhog Day onboarding is authored). After that, organic interaction determines the pace. The prerequisite fires when both the scripted first meeting has happened AND enough subsequent interaction has accumulated. + +The reason this matters for my domain: if Phase Zero is a designed arc (Option A), I can author specific warmth lines tied to specific beats. If it's an emergent threshold (Option B), I need to author lines that can fire in any order across any number of sessions, which means a much larger pool with weaker specific emotional moments. Option C might be the right balance but needs a design decision before content is written. + +--- + +## My Single Most Important Recommendation + +**Design Phase Zero and "First Day" (Ozzie's wow moment #1) as the same thing.** + +Ozzie's minimum viable "First Day" moment: supervisor NPC, tool activation, first assignment acknowledgment. My Phase Zero: meet Kael, run clean jobs, see Naia, build warmth before the arc begins. + +These are not two different design problems. They are the same designed beat described from two different perspectives — Ozzie from experience design, me from narrative arc structure. The first session of a career bookmark should simultaneously: +- Give the player a role (First Day: I belong somewhere) +- Establish the FRIEND figure (Phase Zero: this is the person whose wellbeing I'll be responsible for) +- Teach the career tools organically (Gestalt's diegetic onboarding) +- Deliver place identity through being there, not instruction (Miri's worldbuilding-delivery-as-onboarding) + +One designed beat. Four payoffs. This is where the v0.2 first-30-minutes beat sheet should begin. + +The honest truth is: if we get the first day right — belonging established, tools activated, one specific person made real — everything else in my domain follows. The moral arc can't crack something the player doesn't have. But if they have it, the cracking lands. diff --git a/docs/workshops/wheres-the-fun/round4-tyre.md b/docs/workshops/wheres-the-fun/round4-tyre.md new file mode 100644 index 000000000..b26d4c922 --- /dev/null +++ b/docs/workshops/wheres-the-fun/round4-tyre.md @@ -0,0 +1,175 @@ +# Round 4: Tyre — Cross-Review and Refinement Questions + +**Workshop:** Where's the Fun? | **Round:** 4 (Synthesis) | **Agent:** Tyre (Technical Architect) + +--- + +## Reactions to Other Proposals + +### The consensus is real and architecturally sound + +All 9 agents converged on the same structural conclusion: the simulation engine is correct, the information layer is missing, NPC legibility is the prerequisite. I don't need to re-argue this. What I want to do is pressure-test the *scope* of what's being proposed, because the interview supplement added significant new systems. + +### Gestalt's VerbPriorityProfile — Easy, do it + +The `VerbPriorityProfile` per career is a clean design and a moderate server change. One new field on the character entity, one refactor to the priority sort step in the `verbs[]` computation. No client changes. This is one of those elegant solutions where a single design decision (career-aware verb defaults) teaches the player their role without a single line of tutorial text. I'd put this in Phase 1 alongside NPC legibility. ~1 week server work. + +### Ozzie's redesigned wow moments — Architecturally compatible + +The 6 new moments as emergent thresholds rather than timed beats align perfectly with the life-sim engine. Most of them are queries over existing data: +- "First Day" = onboarding sequence completion event +- "The Character's Instinct" = career-tagged monologue triggers (existing system, new content) +- "The Consequence" = knowledge graph tracking a causal chain back to a player decision (new: needs a `CauseChain` or `DecisionLog` component — we already have CauseChain in the testability architecture, D-030) +- "The Enemy" = relationship state degradation + faction hostility (existing relationship system, needs faction tracking) +- "The Asymmetric Lens" = career-filtered ticker (client-side content rendering, existing snapshot pipeline) +- "The Ownership Moment" = property/asset system (new system — moderate) + +The consequence moment is the most architecturally interesting. We already designed CauseChain as a production component for monologue provenance (D-030 sub-decision 4). Extending it to track "this happened because of your decision 3 hours ago" is a natural growth of that system. *That's actually easier than it sounds.* + +### Paula's Phase Zero — Right call, content-dependent + +The earned comfort period before the moral arc can fire is the correct design. Architecturally, it's a prerequisite gate: `smuggler.phase_zero_complete` must be true before Phase 1-to-2 transition gates activate. This is a knowledge graph flag check — trivial to implement. The hard part is authoring the Phase Zero content (warmth-establishing lines, clean job sequences, Kael/Naia relationship beats). That's Mellanie and Paula's domain. + +### Gore's endgame architecture room — Answering Q-WTF-007 + +Gore flagged that v0.2 architecture decisions must not foreclose the legacy/transcendence endgame path. Let me be direct: **the current architecture already leaves this room open.** + +- The transhumanist ladder (baseline → Higher → ANA) maps to the simulation tier system. A Higher character has expanded perception modes (D-017 already designed for this — perception modes as character build). An ANA character operates at a different simulation layer entirely. The ObserverSnapshot's variable shape accommodates this — different builds, different snapshots. +- Legacy (property, relationships, institutions) maps to knowledge graph + asset ownership + relationship state. All persistent, all serializable, all queryable at endgame. +- Civilizational crisis events are storyteller-driven escalation — the Rimworld model (D-023) already accounts for macro-scale pressure injection. + +**No v0.2 decisions I'm proposing foreclose Gore's endgame.** The skill system (if we build it for CK3-style creation) naturally extends toward Higher capabilities. The asset system naturally extends toward legacy weight. The only thing I'd flag: if we design the skill system now, include a `skill_ceiling` concept that can be raised by transhumanist upgrades later. Don't hardcode human-baseline maximums. + +### Miri and Araminta's dependency chain — Needs mapping + +Qatux correctly identified that my Phase 1 roadmap (NPC legibility in snapshot) doesn't account for the authored content and visual design work that Miri and Araminta describe. Let me be explicit about the dependency: + +``` +Server: NPC legibility data in ObserverSnapshot + ↑ needs: Miri's archetype definitions (what types exist) + ↓ feeds: Araminta's visual rendering (palette per type, behavioral indicators) + ↓ feeds: Client UI (nameplate rendering, activity labels) + +These three tracks can run in parallel if Miri's archetype definitions +come first (design doc, ~1 week) and the server/client/visual work +branches from there. +``` + +The bottleneck is Miri's archetype definition document — everything else fans out from it. + +--- + +## The Interview Supplement Changes My Estimates + +*cracks knuckles* — Let me be honest about what the supplement means technically. + +My Round 3 proposal estimated 4-6 sprints to a playable life-sim vertical slice. The supplement adds three major new systems I didn't account for: + +1. **CK3-style character creation** (skills, culture, family, budget, console cheats) +2. **Three coexisting career models** (Active workplace, WFH/remote, Gig/freelance) +3. **World-first generator pipeline** (uncaring world runs before authored content enters) + +Each of these is substantial. The character creation system alone is probably 3-4 weeks of server work (skill definitions, budget system, starting state derivation). Three career models means three different gameplay loop implementations. And the generator pipeline — if it means procedural world generation — is a multi-sprint endeavor that the Generator Architecture workshop scoped with 14 D-records. + +**My revised estimate: 6-10 sprints if we try to build all of it. That's too long without a playable proof point.** + +This is where the questions for Jeroen become critical. + +--- + +## Questions for Jeroen + +### Question 1: What's the minimum viable "uncaring world"? + +The two-phase approach is the right order — world runs first, authored content enters second. But "the uncaring world generating properly" is a spectrum: + +- **Tier A (hand-built):** Sova Transit as a hand-authored map with NPC routines, an economy tick (wages, rents, prices), and the simulation running. NPCs go to work, go to the bar, go home. Money flows. The world ticks. No procedural generation — same map every time, different world-state variables at start. + +- **Tier B (template-generated):** The Generator Architecture pipeline produces Sova Transit from templates — functional clusters assembled procedurally, NPCs populated by archetype rules, economy seeded from district parameters. Different layout each time but same district character. + +- **Tier C (fully procedural):** The full Cities Skylines top-down pipeline. Geography → infrastructure → zones → population → routines. A new world every game. + +Tier A is 2-3 sprints. Tier B is 4-6 sprints. Tier C is 8-12+ sprints. + +**The question:** For the v0.2 vertical slice — the proof that the life sim works — can we use Tier A? A hand-built Sova Transit with authored NPC routines and a functioning economy, running as the "uncaring world" substrate that authored content (job onboarding, FRIEND arcs, missions) gets injected into? Or does v0.2 need to demonstrate procedural generation to prove the vision? + +**Why this matters:** Tier A lets us prove the life-sim loop in 2-3 sprints of world work + 2-3 sprints of player tools and career systems. Tier B/C pushes the first playable proof to sprint 30+. If the life-sim loop doesn't work, we want to find that out on a hand-built world in sprint 28, not on a procedural one in sprint 35. + +### Question 2: Can v0.2 character creation be "pick skills + pick bookmark" and defer family/culture/religion? + +CK3 character creation is a deep system: family, religion, culture, skill selection, skill levels, point budget, console cheats. All of that is great for the full game. But for the v0.2 vertical slice — the proof that the life sim is fun — how much of that is needed? + +I see a scope spectrum: + +- **Minimal:** Pick a career bookmark. Get a preset character with fixed skills appropriate to the career. No customization. The bookmark IS the character creation. +- **Moderate:** Pick a career bookmark + allocate skill points from a budget. Skills are proficiencies (shooting, hacking, social, mechanical) that gate tool effectiveness and verb success rates. No family/culture/religion yet. +- **Full CK3:** Family tree, culture selection (affects dialogue options and NPC reactions), religion (affects faction relationships), full skill allocation with budget and console cheats. + +Minimal is ~1 week. Moderate is ~3 weeks. Full CK3 is ~6-8 weeks plus content for every culture/religion/family combination. + +**The question:** Which tier proves the life sim? My instinct says Moderate — skill allocation gives the player agency in who they are without the content dependency of culture/religion. But if family and culture are load-bearing for "who am I in this world" — if the life sim doesn't feel like a life without them — then we need to know that now. + +### Question 3: Does v0.2 need all three career models, or can one model prove the concept? + +The interview supplement describes three career models: Active (workplace is gameplay), WFH/Remote (portable income while exploring), and Gig/Freelance (pick jobs from a board). These are three different gameplay loops, each requiring different server systems: + +- **Active** needs: workplace location, shift system, in-job minigames/verbs (cooking, patrolling, processing), supervisor relationship, promotion path +- **WFH/Remote** needs: portable task system (hacking contracts, writing commissions), income-while-exploring model, quality/reputation tracking +- **Gig/Freelance** needs: job board system, contract lifecycle (accept → execute → outcome → payment), varying difficulty/reward, reputation with contractors + +Building all three for v0.2 is ~8-10 weeks of server work. + +**The question:** Can we prove the life sim with one career model and one bookmark? If so, which model best demonstrates the vision? My technical recommendation: **Gig/Freelance** — it's the most gameplay-rich (the player makes active choices about which jobs to take), it naturally teaches the world (jobs take you to different locations), and the job-board system is reusable infrastructure that Active and WFH can extend from. But "Active" (workplace as gameplay) might be more immediately visceral — you're a cook at the bar, and the cooking IS the game. + +Alternatively: one bookmark that blends models. The smuggler bookmark starts Active (onboarding at the dock), becomes Gig (taking smuggling runs from a board), and can eventually go WFH (remote coordination via insert). This tests all three models through one career arc without building three separate systems from scratch. + +--- + +## Agreements Across Domains + +| Agreement | My assessment | +|-----------|--------------| +| NPC legibility first | Correct. Cheapest, highest impact. Unblocks everything. | +| Monologue demoted to supplementary | Correct. Reduces monologue system complexity, increases overall clarity. | +| Diegetic tool suite as core | Correct. This is the bulk of the new work. | +| VerbPriorityProfile per career | Correct. Elegant, moderate effort, high teaching value. | +| Phase Zero before moral arcs | Correct. Trivial to gate architecturally; content-dependent. | +| Visual hierarchy regardless of reframe | Correct. Araminta's work is prerequisite infrastructure. | + +## Conflicts That Need Resolution + +**1. The career bookmark for v0.2 — law enforcement vs smuggler.** + +Ozzie says law enforcement (most legible first run). I said smuggler in Round 3 (most content exists). The interview supplement changes my thinking. If we're building one bookmark to prove the life sim, the career model matters more than the content stockpile. The smuggler's existing content was authored for a detective-game context — it may need significant rework for the life-sim framing anyway. Law enforcement has the advantage of clear institutional onboarding (Commission orientation, insert activation, weapon qualification) which naturally teaches the world. + +**My revised recommendation: law enforcement.** The institutional onboarding is the most natural diegetic tutorial. The content gap is real but bounded — the onboarding sequence needs to be authored from scratch regardless of which bookmark we pick. + +**2. Mission system design — the supplement reframes this significantly.** + +My Round 3 proposal included a `MissionState` + `ConsequenceEngine` with objective tracking and outcome spectrums. The interview supplement says missions are "emergent things that happen in the context of what the player chooses to do." That's less structured than what I proposed. + +The consequence engine still applies. But the formal mission system with `ObjectiveId` tracking may be overbuilding. What Jeroen described is closer to: the player has a job, things happen during the job, outcomes have consequences. The "mission" is the job itself, not a quest tracker. This is simpler to build but harder to design — it requires the consequence engine without the objective structure. + +I'd still want a design workshop for this. But the scope may be smaller than my Round 3 estimate. + +--- + +## My Single Most Important Recommendation + +**Build the smallest possible proof before building the full vision.** + +The interview supplement describes an extraordinary game. CK3 character creation, three career models, procedural world generation, emergent missions with consequence spectrums, transhumanist endgame hooks. I'm genuinely excited by it. But scope-wise, this is 10-15 sprints of new systems before first playtest. + +The v0.1 lesson was: we built too much before testing. We shipped a technically correct engine that wasn't fun because we never tested the fun hypothesis early enough. + +For v0.2, I propose a **proof-of-life sprint** before the full buildout: + +1. Hand-built Sova Transit with NPC routines and economy (Tier A world) +2. One career bookmark (law enforcement), moderate character creation (skills + career) +3. NPC legibility (names, activities, types visible) +4. One diegetic tool (thread tracker or journal) +5. Three days of gameplay: onboarding → first assignment → first consequence + +If that's engaging at the playtest — if the player knows what to do, cares about people, and sees consequences — then we build the full vision on proven ground. If it's not, we've invested 3-4 sprints instead of 10. + +The engine is right. The vision is right. The risk is building the cathedral before testing the foundation. Let's pour the foundation, test it, then build. diff --git a/docs/workshops/wheres-the-fun/round5-araminta.md b/docs/workshops/wheres-the-fun/round5-araminta.md new file mode 100644 index 000000000..4820fadd9 --- /dev/null +++ b/docs/workshops/wheres-the-fun/round5-araminta.md @@ -0,0 +1,467 @@ +# Round 5: Araminta — Visual Design Reaction to 15 Decisions +## Where's the Fun? Workshop | 2026-03-05 + +--- + +## Overview + +The 15 decisions land well conceptually. The generator-first approach, full character customization, and generated NPCs are all correct directions. But three of these decisions create compounding visual design problems that aren't solved by "we'll use outline/highlight." This round is my full technical accounting of those problems, plus per-decision reactions for completeness. + +--- + +## Decision-by-Decision Reactions + +### Decision 1: Proof-of-life = generator + graphics, not hand-built slice + +**Strong agreement — and it fundamentally changes my design task.** + +I was planning to design "Sova Transit's visual identity." That's the wrong output. The generator doesn't produce Sova Transit specifically — it produces locations with characteristics, and Sova Transit is one instance of a location that has certain zone types. My output needs to be a **visual grammar the generator can apply**: rules that say "a logistics zone has this color temperature, this prop vocabulary, these NPC archetype expectations." Sova Transit is the PROOF of that grammar, not the deliverable. + +This means the style guide I write isn't world art direction for a specific place. It's a parameterized visual grammar with zone type as the input and palette + prop vocabulary + NPC silhouette expectations as the output. The generator reads from this grammar to dress any location it produces. + +Implication for sprint planning: the style guide draft needs to be written BEFORE any sprite or tile work begins, because it defines the rules that all other visual production follows. + +### Decision 2: Skills + bookmark only for character creation (family/culture deferred) + +**Noted with a visual design subtlety.** + +Skills and bookmarks are the creation axes. Culture selection is deferred. But Decision 6 says culture is PRIMARY for voice — culture-driven, job modifies. This creates a gap: if the player doesn't select culture at creation time, where does their cultural identity come from for visual purposes? + +Options: +- **Implicit from starting location**: the generated starting apartment and neighborhood express the cultural context the character is embedded in, even if not selected +- **Bookmark implies culture**: the tycoon bookmark implies a certain cultural-economic position, and the generated world populates the relevant cultural visual context around them +- **Character appearance choices ARE the cultural signal**: the player picks hair, clothing, colors — and those choices encode cultural belonging implicitly, without a culture dropdown + +I'm operating under the assumption that the player's appearance choices (Decision 11) are the cultural expression in character creation, and that the world around them reflects generated cultural context. But this should be confirmed before designing the character creation screen. + +### Decision 3: Religion is NOT a game system + +**Clean scope removal. No visual design implications.** + +### Decision 4: Tycoon is the v0.2 bookmark + +**Sets my visual priority sequence.** + +The tycoon bookmark gets the first career insert design. Visually this means: +- **Insert aesthetic**: financial dashboard. Warmer palette than law enforcement's institutional blue. Market data feeds, asset position summaries, economic opportunity indicators. The Settled Reach equivalent of Bloomberg Terminal meets a person's neural HUD. +- **Tycoon verb visual treatment**: economic verbs (Buy/Invest/Negotiate/Contract) need visual weight that communicates stakes. These aren't investigation verbs that reward patience — they're decision verbs with visible consequence speed. +- **World overlay priority**: property ownership markers, business contact availability indicators, economic zone legibility. + +The tycoon insert will be the template that the other career inserts are designed against. Getting the tycoon design right establishes the grammar that law enforcement and other bookmarks follow in v0.3+. + +### Decision 5: Skills affect outcome, not verb availability + +**Outcome feedback becomes a visual design problem I hadn't fully addressed.** + +If everyone sees the same verbs but skill determines quality of outcome, then the POST-verb feedback (what happens after you act) needs to communicate outcome quality. A skilled negotiation vs a fumbled one produce different results — that difference needs visual representation. + +Options for outcome quality signal: +- **Feedback animation quality**: a skilled action resolves cleanly; a poor skill action has a visual stumble (brief flash of incorrect/off-center feedback) +- **Outcome descriptor in the signal**: the ambient tier feedback line says "Negotiation: strong position" vs "Negotiation: they noticed your hesitation" — the text carries the quality signal +- **NPC behavioral response**: the NPC's behavioral state indicator shifts based on outcome quality — still moves to neutral after a stumble, but to warm after a confident interaction + +I favor the third option for life-sim feel — the feedback is social/behavioral, not scored. You read the room, not a number. But this needs NPC behavioral state indicators to be designed and live in the client before it works. + +### Decision 6: Culture-driven voice, job modifies + +**Sets the NPC visual character design priority.** + +NPCs are generated with cultural identity as the primary dimension, job as modifier. Visually this means: +- **Silhouette/posture = role/job** (the dock worker stands differently than the business contact) +- **Palette and detail = culture** (the same dock worker from culture A vs culture B has the same body posture but different clothing palette and cultural detail) + +The risk: if I design the archetype system with ONLY role as the organizing axis, I'll produce archetypes that look culturally flat — everyone's a generic type, not a person from somewhere. The visual grammar needs culture as a modifier layer on top of role as the structural layer. + +This is a component-asset design question (see Deep Dive 2 below). + +### Decision 7: ALL NPCs are generated, no named characters + +**The biggest visual design scope change in the 15 decisions.** + +I was planning to design visual identities for Kael, Naia, Maret. Those characters don't exist. The visual system needs to produce emergent character identity from generated components without hand-authored personality designs. + +The key insight from Jeroen's answer: "The Sims and Rimworld are perfectly capable of generating interesting characters that come alive and create attachments." The Sims achieves this through: +1. Distinct visual appearance (generated unique combination of features) +2. Behavior that expresses personality consistently +3. Player investment through relationship mechanics + +The visual design task: create a component-assembly system that reliably produces characters that feel DISTINCT from each other (not all looking like variants of the same template) and LEGIBLE in role/relationship (not visually indistinguishable chaos). + +See Deep Dive 2 for the full technical breakdown. + +### Decision 8: Generative AI for NPC content templating + +**Noted — not directly a visual design task, but has one implication.** + +The AI templating is for content (dialogue, tone). Not for sprite generation. My component-asset system produces sprites; AI assists with the personality/dialogue layer. The two pipelines are parallel, not coupled. + +The implication: eventually, the AI-generated personality of an NPC (their cultural tone, their speech patterns) should INFORM what kind of visual variation they receive from the component system. A character generated as "reserved, formal, high-status" should visually read as such — formal clothing set, upright posture, muted palette. This mapping from personality/culture dimensions to visual components needs to be in the style guide. + +### Decision 9: Possible in-game ollama for live NPC dialogue + +**Deferred — one future visual design consideration.** + +If NPCs have live AI dialogue, there may be a moment where the NPC is "thinking" (generating a response). A visual treatment for this state — subtle, not breaking the diegetic fiction — would be needed. Something like a brief attentive posture hold, a minimal insert-style processing indicator in the dialogue panel. Not a spinner. Something that feels like a person collecting their thoughts rather than a machine processing. + +Flagging for later; no immediate visual design action needed. + +### Decision 10: Quietly responsive world, gradient of caring by social proximity + +**The behavioral state indicator system needs a relationship axis.** + +My Round 3 behavioral state indicators were designed for three states: stress, suspicion, routine comfort. That's a world-state axis. Decision 10 adds a RELATIONSHIP axis: this NPC notices you, cares about you to varying degrees. + +The relationship gradient (stranger → acquaintance → colleague → friend) needs visual expression. My proposal: + +| Relationship level | Behavioral indicator treatment | +|---|---| +| Stranger | No indicator (world doesn't know you exist) | +| Acquaintance | Ambient tier indicator on close approach (grey, low-weight) | +| Colleague/neighbor | Relevant tier indicator when active (amber, present but not demanding) | +| Primary social contact | Relationship state visible (warm/stressed/concerned — behavioral reads now include relationship reads) | +| Hostile | Danger tier signal (red, avoiding/tracking you) | + +This is still consistent with the three-tier system. The tier isn't just about urgency — it also maps to social proximity. Strangers live in ambient-or-invisible. Friends are legible at the relevant tier unless they're in crisis, which promotes to danger. + +### Decision 11: Full character customization (hair, clothing, colors) + +**The hardest unsolved visual design problem. Deep dive below.** + +Acknowledged as a real problem. Outline/highlight is the proposed solution. I have significant concerns about how this plays out at tile scale. See Deep Dive 1 for the full breakdown. + +### Decision 12: Setting delivery: both layers (visual + insert) + +**Confirms the parallel production structure.** + +The visual layer shows place identity through palette, props, and NPC behavior. The insert layer names and contextualizes. These need to be designed together — the insert copy for a zone should reference the VISUAL elements the player is already seeing, not introduce information the world hasn't shown. "The docks" is named by the insert after the player has already read the space as industrial/logistical from its visual grammar. + +Practical implication: Mellanie and I need to co-draft the tycoon insert copy AFTER the zone visual grammar is established, so the copy can reference the visual context accurately. + +### Decision 13: First Settled Reach moment: apartment + insert activation + +**Two designed visual sequences, each deserving their own spec.** + +**The Apartment Wake-Up:** + +The apartment is auto-generated to reflect economic starting position. For the tycoon bookmark with a decent skill budget: a mid-to-upper apartment. But "mid-to-upper" needs visual specifications the generator can apply: + +- **Wealthy apartment signals**: larger window viewport (more of the exterior visible), furniture density higher, decor items present (art, plants, personal objects), cleaner color palette (less wear-and-tear textures) +- **Modest apartment signals**: tighter viewport, fewer furniture pieces, more utilitarian props (a single chair, a small screen), slightly worn palette + +These aren't just size differences — they're VISUAL GRAMMAR signals that communicate economic status without explicit text labels. A player who picked a skill-heavy starting budget (less starting capital) should wake up in a space that visually reads as "I'm building toward something, not there yet." + +The apartment is also the player's first glimpse of the Settled Reach's visual technology — insert hardware on the wall or bedside, ambient smart-building elements, the window showing the generated exterior. The Settled Reach reads as advanced-but-human through the apartment's tech-texture, not through a lore drop. + +**Insert Activation:** + +The neural insert powering on is a DESIGNED BEAT. Visually I'd propose: +1. Apartment view is clean (no HUD overlay) +2. Insert activates: a subtle warmth blooms across the vision periphery (edge vignette in insert accent color, very brief) +3. UI elements phase in sequentially: world overlay layer first (you can now "read" zone labels), then the financial feeds, then notification panel — each appearing as the insert calibrates +4. The tycoon insert's dashboard assembles itself — empty at first, populating with starting data + +This sequence teaches the player what the tycoon insert shows WITHOUT a tutorial. They watch it come online and see what it surfaces. The insert IS the onboarding. + +The visual register for insert activation should feel intimate and sensory — not technical UI appearing. The neural connection is a felt experience, not just a screen appearing. Blur-to-clarity transition. Sound design partners with this heavily (Gore/GORE's domain but the visual sequence needs to be audio-aware). + +### Decision 14: Groundhog Day alarm clock homage + +**Specified in audio, implied visually.** + +The *click* pa-pa pa-pa is audio. The visual analog: the alarm clock device in the apartment (whatever form that takes in the Settled Reach) transitions from "idle/sleeping" state to "active" state as the sound fires. Then cut short — the device returns to idle as the player is "already awake" before the full sequence plays out. + +The cut-short visual: the device's glow/indicator starts its wake-up pattern, then snaps off as the player's perspective says "I'm already up." This is subtle and fast. It shouldn't demand the player's attention — it's a background wink, not a front-and-center animation. + +The "first game day only" constraint is important. On subsequent days, the alarm is ordinary (different, understated). This wink exists to establish the tone of the whole game from the very first second: this world is lived-in and wry, not earnest and serious. + +### Decision 15: Player choices ARE the content (Rimworld model) + +**Changes the insert's visual job description.** + +The insert can't be organized as a quest tracker — there are no quests. It's organized as a WINDOW ONTO OPPORTUNITY. The tycoon insert shows: +- Current market positions (what's moving, what's stable, what's vulnerable) +- Social contacts and their availability status +- Property and business asset summary (what you own, what state it's in) +- Economic events and news (filtered through the tycoon's knowledge/network) + +This is a DASHBOARD, not a briefing. The visual design challenge: a dashboard can easily become overwhelming (too much data) or sparse (nothing to read). The 3-tier signal hierarchy applies here too. The insert surfaces: +- **Danger tier**: something you own or control is threatened, a deal is going bad +- **Relevant tier**: a contact is available, a market opportunity is open, a decision is pending +- **Ambient tier**: background world data, the economic texture the player reads over time + +The dashboard design should feel like the player is reading the temperature of their situation at a glance. Not: "here's your objective." More: "here's the state of your world; what do you want to engage with?" + +--- + +## Deep Dive 1: Full Customization at Tile Scale + +### The Core Tension + +The creation screen is an emotional investment moment. The player has chosen their character's hair color, clothing style, palette. They've invested in who this person IS visually. Then they enter the game — and their character is a 16-32 pixel sprite that needs to be readable at a glance during active gameplay. + +These two modes of seeing the same character have fundamentally different readability requirements. Creation-screen fidelity and tile-scale gameplay fidelity are not on the same spectrum — they're different problems. + +### The Outline/Highlight Approach — What It Actually Solves + +Outline/highlight solves one specific problem: **separating the player character from the background and other NPCs at tile scale.** A consistent 2px outline in a high-contrast color means the player's character doesn't visually merge with the tile floor or NPC crowd. + +What it does NOT solve: +- Whether the player's customized hair color reads differently from an NPC's hair color at 16 pixels +- Whether a player who chose "red jacket" looks meaningfully different from a player who chose "dark jacket" in a crowd +- Whether the emotional investment of the creation screen translates to a feeling of "that's MY character" during gameplay + +### The Actual Solution: Identity Tokens, Not Pixel-Fidelity + +The answer is not to make tile sprites represent creation choices at high fidelity. The answer is to make the **impression** of creation choices survive into tile scale. + +What survives tile-scale reduction: +- **Color temperature** (warm vs cool palette) +- **Value contrast** (light vs dark clothing overall) +- **Silhouette** (slim vs broad, fitted vs loose clothing) +- **Hair color read** (dark/light/vivid — three distinguishable reads at 16px) + +What doesn't survive: +- Specific haircut shapes +- Clothing detail (pockets, collars, material texture) +- Facial features + +**Proposed approach: Creation choices map to tile-scale token sets.** + +The creation screen offers full-fidelity selection. Each selection maps to a tile-scale token that represents the impression of that choice: + +- "Deep red jacket" → tile sprite uses the warm-red jacket asset in the relevant body type layer +- "Ashen short hair" → tile sprite uses the dark-hair-short asset in the hair layer +- "Lean build" → tile sprite uses the slim silhouette body layer + +The player sees their character in full fidelity during the creation screen and in portrait/close-up contexts. During tile-scale gameplay, they see an impressionistic representation that carries the COLOR TEMPERATURE and SILHOUETTE of their choices without the detail. + +This is what The Sims does: the Sim looks like their creation at full zoom, like a colored impression at distant zoom. The player's brain fills in the detail because they know what they chose. + +**Identity lock: the player character outline color is theirs.** + +The player's character has a consistent outline color that is chosen during creation or assigned at generation and never duplicated in the local scene. This is not the avatar's clothing — it's a subtle but consistent visual marker. During gameplay: you always know which sprite is yours because of the outline treatment. No cognitive load. + +### The NPC Variant + +NPCs are generated. They have the same component-assembly system but no specific player investment in their appearance. Their tile-scale impression needs to communicate role first (archetype silhouette), culture-context second (palette), and individual distinctness third (within-type variation). + +The risk: if NPC generation produces too much within-type variation, the archetypes stop reading. If it produces too little, everyone in the same role looks identical. + +**Proposed constraint: archetype silhouette is fixed per role. Palette is variable within cultural range. Detail (hair, accessory) is variable within cultural range.** + +A dock worker's POSTURE and BODY TYPE silhouette doesn't vary. What varies is the color of their work gear (culture-influenced) and their hair. At tile scale, they still read immediately as a dock worker — but they look like THEIR dock worker from wherever they're from. + +### What Needs Confirmed Before I Can Design This + +1. **Creation screen fidelity target**: is the creation screen a full-character render (like CK3's portrait) or a top-down tile-scale preview? If CK3-style, I'm designing two character representations — portrait and tile. If tile-only, the creation screen shows the impressionistic version and full fidelity is never the goal. +2. **Component granularity**: how many distinct options does the player choose from in each category? The granularity determines how many tile-scale token assets I need to produce. 10 hair colors × 8 clothing styles × 4 silhouette types = 320 sprite combinations per archetype. At tycoon-only scope, this is manageable. At all-careers scope, it scales fast. + +--- + +## Deep Dive 2: Generated NPC Visual Variety Without Visual Noise + +### The Problem + +All NPCs are generated. The generator produces: +- Role (dock worker, colleague, business contact, neighbor) +- Cultural background (which culture's visual grammar applies) +- Personality dimensions (reserved vs. expressive, formal vs. casual, high-status vs. low-status) +- Relationship to the player (stranger, acquaintance, contact, rival) + +The visual system needs to express all of these dimensions while keeping any NPC legible at a glance for ROLE. A player looking at a crowd needs to instantly parse: "dock workers over there, two business contacts near the entrance, a supervisor at the desk." They don't need to consciously read this — the visual grammar communicates it before the player's conscious attention engages. + +### The Legibility Hierarchy + +**Rule: role legibility must survive ALL cultural variation.** Two dock workers from different cultures must both instantly read as dock workers. Cultural variation lives in dimensions that don't obscure the role signal. + +Dimensions that carry ROLE signal (fixed per role, not culturally variable): +- Body posture / silhouette shape +- Clothing TYPE (working gear vs. business wear vs. casual) +- Behavioral default (what they're doing when idle) + +Dimensions that carry CULTURE signal (variable within cultural grammar): +- Color palette of clothing (same clothing type, different cultural palette) +- Cultural detail markers (cultural accessories, style flourishes) +- Hair treatment (cultural norms vary) + +Dimensions that carry INDIVIDUAL signal (variable within cultural range): +- Specific palette choice within cultural range +- Hair color/type selection +- Minor accessory presence + +This hierarchy means: at tile scale, you read posture+clothing type → role. Then you read palette → cultural context. Then, if you're paying close attention, you read individual variation. Role is always the first signal; individual distinctness is the last signal. + +### The Component System Architecture + +Each NPC sprite is assembled from four layers: + +1. **Body layer** (fixed per role category): 3-5 distinct role silhouettes (service worker, professional, manual worker, authority figure, specialist/technical). This layer doesn't vary within a role. + +2. **Clothing layer** (variable per cultural grammar): each role has a clothing type (work gear, business wear, etc.) with 4-8 cultural palette variants. The clothing type is fixed; the palette is culture-selected. + +3. **Hair layer** (variable within cultural range): 3-4 hair type options per body layer, palette variants within cultural range. + +4. **Detail layer** (optional, culturally significant markers): small cultural detail elements (a specific type of jacket closure, a cultural marking, an accessory associated with cultural background). Present or absent based on cultural generation rules. + +**Total sprites at v0.2 scope (tycoon bookmark, 3 role types, 3 cultural variants, 3 hair options):** +3 body types × 3 cultural clothing palettes × 3 hair options × 2 detail states = 54 tile sprites per direction × 4 directions = 216 total sprite assets for NPC variety. This is a manageable production run. Scales with role types and cultural breadth as the game grows. + +### The Distinctness Problem in Crowds + +Even with the hierarchy above, a crowd of same-role NPCs (all dock workers) risks visual repetition that undermines the "these are distinct people" reading. The Sims solves this with aggressive HAIR COLOR variation — even if body and clothing read similarly, distinct hair colors make individuals visually separable at a glance. + +**Proposed rule: within any scene, no two NPCs share the same combination of clothing palette + hair color.** The generator tracks this during scene population and prevents duplicate combinations. Players can learn to tell individuals apart by their color combination even before knowing their name, which is how you start to form "that's the one who..." recognition before a name is revealed. + +This is the visual equivalent of the knowledge-gated name reveal: you recognize the COMBINATION before you know the person. + +--- + +## Deep Dive 3: Auto-Generated Apartment Visual Variety + +### The Problem + +The apartment reflects economic starting position. This is meaningful as a first impression — your apartment tells you something true about who your character is starting as. But "wealthy" and "poor" are mechanical designations; the visual system needs to translate them into spatial and atmospheric signals without explicit labels. + +### The Visual Grammar of Economic Status in the Settled Reach + +What communicates wealth in a near-future working-class-to-middle-class range (the likely tycoon starting range)? + +**Wealthy signals (in Settled Reach aesthetic):** +- Larger viewport — the window shows more of the exterior, implying more floor space +- Furniture density — more objects, each with finer detail +- Color palette — cleaner, less worn. Neutral-warm tones, intentional decor choices +- Technology presence — visible insert station, ambient-smart-building elements (lighting that responds, wall panels that indicate rather than just existing) +- View quality — the window looks out on a better slice of the generated exterior (elevated, park-adjacent, not directly onto another building face) + +**Modest signals:** +- Tighter viewport — implied smaller space +- Fewer furniture pieces — one chair, a bed, a screen. Utilitarian +- Color palette — slightly warmer from use, worn textures visible +- Technology presence — insert hardware present (everyone has inserts) but less polished; the wall unit is clearly third-party or older model +- View quality — street level, another building face close, less natural light implied + +**Critical rule: both apartments must feel INHABITED, not empty or depressing.** This is not Kenshi-indifference — the world quietly cares. Even a modest apartment has one personal object that communicates this is someone's home. A cheap screen with something playing, a plant that's doing okay, a jacket hung by the door. The emotional register is: "this person is making something of what they have" across the full wealth range, not "poor people have sad apartments." + +### Generator Parameters + +For the generator to produce apartments with these visual reads, it needs parameters it can read from the character's starting state: + +- `wealth_tier` (1-5 or similar) → selects furniture density level + palette set + viewport size +- `cultural_context` → selects which cultural visual grammar applies to the apartment's design language (furniture style, decor type, personal objects) +- `bookmark` → potentially adds career-specific personal objects (a tycoon even at starting wealth has one business-related item visible — a small screen showing market data, a leather planner-equivalent, something that says "this person thinks about money") + +The personal objects are the richest legibility signal. They're small, but they're the thing that makes the apartment feel like a CHARACTER'S APARTMENT rather than a generated space that happens to belong to whoever plays it. + +--- + +## Cross-References to Other Agents' Round 4 Work + +### Miri (Round 4) — Zone Identity Spec is My Direct Upstream Dependency + +Miri identifies the same chain I identified in Round 3: Miri writes the zone identity spec → +Araminta produces tile palettes → Tyre exposes zone-type data in snapshot → generator applies +visual grammar. Miri calls the zone identity spec "the first worldbuilding deliverable in the +v0.2 roadmap." I agree completely. I cannot author the generative zone visual grammar (Decision +1's actual deliverable) until Miri's spec defines what zone types mean in the Settled Reach's +social vocabulary. + +The apartment visual grammar is a subset of this: "residential zone, lower economic tier" is +a zone type with a social meaning. Miri and I should design these together — one document with +a worldbuilding section and a visual expression section, not two documents that need reconciling. + +### Tyre (Round 4) — Archetype Taxonomy Is a Joint Design Task + +Tyre's dependency mapping shows the server-side data (NPC legibility in ObserverSnapshot) can +only expose what Miri defines as existing archetype types. My visual system can only render what +the snapshot exposes. These three tracks fan out from Miri's archetype definitions. + +Critically: Tyre's dependency chart shows "Miri's archetype definitions (what types exist)" as +the bottleneck that everything fans out from. But the visual grammar needs to INFORM those +definitions too — some archetype distinctions only exist because there's a meaningful visual +difference. If two archetypes look identical at tile scale, they should be one archetype. The +design flow needs to be: Miri/Araminta co-design the archetype taxonomy → Tyre implements it in +the snapshot. Not: Miri designs, Tyre implements, Araminta adapts. + +### Gestalt (Round 4) — Verb Prompt Icon Set Confirmed In Scope + +Gestalt's VerbPriorityProfile visual surface comment from Round 3 is still unaddressed in Round +4. I flagged it in my Round 4 output and it's now formally in my deliverable list (item 8: +"Verb prompt icon set per career"). The icon set is small work with high consistency value — +the icon inside the [E] bracket signals career register before any text appears. This is the +visual layer underneath Gestalt's career-aware verb ordering. + +### Ozzie (Round 4) — Wow Moments Still My Visual Spec List, Now Tycoon-Filtered + +Ozzie's six redesigned wow moments remain my visual specification list. They're now read through +the tycoon lens: +- "First Day" = arriving at the tycoon's business for the first time — the Ownership Moment's + antecedent +- "The Ownership Moment" = the tycoon's primary emotional peak — property tile with ownership + indicator, business health state visible +- "The Consequence" = an economic decision's downstream effect becomes visible on the insert + (a deal you made three days ago collapsed someone else's margin) +- "The Asymmetric Lens" = the market data ticker filtered through the tycoon's network and + investment position — what they see that others don't + +These are concrete design targets. I need Ozzie's final wow moment list (post-Round 5) before +the insert grammar document can be written. + +### Paula + Gore (Round 4) — Phase Zero and Consequence Require NPC Legibility First + +Both Paula's Phase Zero warmth model and Gore's consequence-as-theme require the player to +emotionally invest in generated NPCs before any arc can fire. Both are gated on the visual +system producing legible, distinct, relationship-aware NPC representations. My deliverable +ordering (NPC legibility rules first, everything downstream of it) is the visual prerequisite +for their content to land. + +### Mellanie (Round 4) — Tycoon Insert Grammar Is a Joint Deliverable + +My Round 4 recommendation stands: the career insert grammar document should be co-authored +(visual register column: me; voice register column: Mellanie; one per career). With tycoon +confirmed as the v0.2 bookmark, the tycoon insert grammar document is the first output. The +tycoon's visual register (financial dashboard, market data, amber-warm palette) needs to be +designed in the same session as the tycoon's voice register. If I design first and Mellanie +adapts, or vice versa, they'll feel like two designers on the same brief who never spoke. +Short joint session, single document, one visual register and one voice register per section. + +--- + +## Questions for Jeroen (Visual Design Near-Misses) + +### Q1: Creation screen fidelity — portrait or tile preview? + +The emotional investment of character creation depends on the player seeing their character at a fidelity that makes them care. At tile scale (16-32px), no character looks emotionally distinctive enough to invest in. CK3 uses full portraits. The Sims uses a fully-rendered 3D character viewer. + +The near-miss: we design a creation screen that shows a tile-scale preview of the character — a technically accurate view of what they'll look like in gameplay — and the player can't tell the difference between their customization choices because everything is 20 pixels tall. The identity investment moment fails because the creation screen didn't make them care. + +**Question**: Is the character creation screen a full-fidelity portrait/render moment, or is it tile-scale? And connected: when the player sees their character in the apartment wake-up sequence (their first in-world moment), is that a close-up portrait view or a tile-scale view? The answer determines whether I'm designing two character art styles or one. + +### Q2: How many cultures exist in the generated world, and what's their visual distinguishability priority? + +Culture is primary for voice (Decision 6). Culture presumably affects apartment design, NPC appearance, and world texture. But we deferred culture selection from character creation — which implies culture is a WORLD property (what culture is this location's dominant culture) rather than a character property the player selects. + +The near-miss: I design 2-3 cultural visual variants (enough for a prototype), and the world generates 6-8 cultural contexts because the Settled Reach's setting has established several distinct cultures. I've under-built the system; the generator exposes how incomplete the component library is. + +**Question**: How many distinct cultural visual contexts should I be designing for at v0.2 scope? And is cultural visual grammar something the world generator applies to LOCATIONS (this location has a dominant cultural context) or to INDIVIDUAL NPCS (each NPC has a cultural background that may differ from the location's dominant culture)? The answer changes the combinatoric complexity significantly. + +### Q3: The apartment as identity anchor — how personal is it? + +Jeroen described the apartment as part of the "First Settled Reach moment" alongside insert activation. The apartment is auto-generated but should reflect economic position. My deep dive above proposes that bookmark also affects personal objects (a tycoon's apartment has tycoon-relevant items). But how personal is it, really? + +The near-miss: the apartment is generated to reflect wealth level but feels like a hotel room — correctly affluent or modest, but not YOURS. The player doesn't see themselves in it. They wake up in a room that is objectively appropriate but subjectively empty. The identity investment moment of character creation doesn't carry into the apartment because the apartment has no visual relationship to the choices the player made. + +**Question**: Should the character's creation choices (appearance selections, skill emphasis) leave any trace in the apartment's visual generation? For example, a physically-skilled starting character might have equipment visible; a social-skill emphasis might have more relationship memorabilia visible. Or is the apartment purely economic position + cultural context, and the player's specific creation choices aren't expressed there? + +--- + +## My Single Most Important Concern + +**"Outline/highlight solves readability" is a hypothesis, not a solution.** + +Jeroen's answer to the full customization question was: "Readability solved through outline/highlight, not by limiting customization." I understand this and I'm not arguing against full customization. But outline/highlight is a disambiguation tool — it separates figures from backgrounds and from each other. It does not make a 16-pixel sprite communicate "this is the character I spent 15 minutes designing." + +The emotional continuity between creation screen and gameplay is the design problem that needs solving before we commit to what "full customization" means in implementation. If the creation screen shows a tile-scale preview, the player knows what they're getting and invests at that level. If the creation screen shows a high-fidelity portrait, the player invests at portrait level and then needs to FIND that investment in gameplay — which means the tile sprite needs to carry the impression of the portrait, not be a separate, lower-fidelity thing. + +I need to know what the creation screen fidelity target is before I can design the character component system with confidence. Everything downstream — how many components, what tile sprite resolution, whether to use portrait close-ups in apartment and dialogue contexts — depends on that answer. + +This isn't blocking sprint work immediately. The zone palette grammar and NPC archetype silhouettes don't depend on it. But before any character creation screen or PC sprite production begins, this question needs a confirmed answer. diff --git a/docs/workshops/wheres-the-fun/round5-gestalt.md b/docs/workshops/wheres-the-fun/round5-gestalt.md new file mode 100644 index 000000000..a1cf8a7a7 --- /dev/null +++ b/docs/workshops/wheres-the-fun/round5-gestalt.md @@ -0,0 +1,492 @@ +# Round 5: GESTALT — Systems Design Reaction to 15 Locked Decisions +## Where's the Fun? Workshop | 2026-03-05 + +**Agent:** GESTALT (Systems Design) +**Based on:** Round 4 interview transcript, interview supplement, all Round 4 agent outputs + +--- + +*Let me break down what these 15 decisions actually mean mechanically. Crack knuckles. Some of these are clean. Some of them open new design work I wasn't expecting. And one pair of decisions is in tension that nobody has flagged yet.* + +--- + +## Part 1: Decision-by-Decision Mechanical Analysis + +### Decision 1 — Proof-of-life = generator + graphics + +This is the single most scope-altering decision in the batch, and I don't think its full weight has landed yet. + +Jeroen explicitly rejected the reasoning that led to v0.1's narrow scope. The proof-of-life is NOT a hand-built Sova Transit with authored NPC routines. It IS the generator running at reasonable scale + legible characters. That means Tyre's Tier A ("hand-built, different map every time but same district character") is off the table. We need at minimum Tier B (template-generated locations, archetypes populated by rules). + +**What this means mechanically:** + +The 14 D-records from the Generator Architecture workshop are now load-bearing for v0.2. The generator isn't a v0.3 feature we're planning toward — it's the v0.2 deliverable. Every other system (NPC legibility, career bookmarks, the apartment) is built ON TOP of the generator, not instead of it. + +This also means Decision 7 (all NPCs generated) and Decision 1 are the same pipeline, not two separate problems. The generator produces: locations → zones → NPC populations. The NPC archetypes fill roles based on zone characteristics. The generator IS the NPC generation system. + +**Implementation risk:** The generator-first approach defers the "is the life-sim fun?" validation. We won't have a playable first-run experience until the generator is producing legible space. If the generator takes 4-6 sprints (Tyre's Tier B estimate), the first "does this feel like a game" moment is sprint 28-30. That's the v0.1 mistake in a different form — building foundation before testing the experience layer. + +**Flag:** This is a real tension between Decision 1 (generator first) and the principle of "smallest possible proof before full vision." I'm not saying Jeroen is wrong — the generator IS the vision, not an optimization. But someone should be tracking when the first playable proof moment arrives in the generator-first roadmap. + +--- + +### Decision 2 — Skills + bookmark only for character creation + +Clean and right for scope. Family/culture/religion deferred. + +**What this means mechanically:** + +The character creation system for v0.2 needs: (1) skill point allocation across proficiencies, (2) career bookmark selection. That's it. + +But Decision 6 (voice is culture-driven) creates a tension I need to flag. See Part 2 below. + +**Skill proficiency list:** The supplement names shooting, social manipulation, hacking, mechanical repair. The tycoon bookmark (Decision 4) probably leans on social manipulation and possibly hacking (for insert operations). Shooting and mechanical repair are secondary. Does "skills + bookmark" mean all four proficiencies are available to allocate regardless of bookmark, or does the bookmark constrain which skills are relevant? This needs a spec. + +--- + +### Decision 3 — Religion is NOT a game system + +Removed. No mechanical footprint. No action required. + +--- + +### Decision 4 — Tycoon is the v0.2 bookmark + +This decision resolves the Tyre/Ozzie dispute about law enforcement vs smuggler by going in a completely different direction. Zero investigation content. Tycoon. + +**What this means mechanically:** + +The entire verb ecosystem I was preparing to spec needs to start from scratch with tycoon verbs. What does a tycoon DO, action-by-action? + +I don't have a spec for this. Nobody does. This is the first concrete design work needed for v0.2. + +My initial tycoon verb map: + +| Context | Primary Verb | Secondary Verbs | VerbPriorityProfile default | +|---------|-------------|-----------------|----------------------------| +| Approaching an NPC (business) | Negotiate | Examine, Talk, Observe | Negotiate > Examine > Talk > Move | +| Approaching an NPC (social) | Talk | Observe, Move | Talk > Observe > Move | +| Approaching a location (economic) | Invest | Inspect, Inquire | Invest > Inspect > Inquire | +| Insert use (remote) | Monitor | Adjust, Transfer | Monitor > Adjust | +| Gig contract (one-off deal) | Accept | Examine, Decline | Accept > Examine > Decline | + +This is a first pass, not a spec. The tycoon VerbPriorityProfile is the first thing I need to design before any downstream work can proceed. + +**Tycoon career model synthesis:** The decision notes that tycoon blends Active (manage business), WFH (remote investments via insert), and Gig (one-off deals). This is the first bookmark that explicitly touches all three career model rhythms. Mellanie's concern about three monologue cadences is therefore relevant to the tycoon specifically — the tycoon isn't one rhythm, it's three rhythms in one career. That's elegant if designed intentionally, but complex if not. + +--- + +### Decision 5 — Skills affect outcome (mostly C) + +My Round 4 Q1 is answered. This is the right answer and I endorse it. + +**What this means mechanically:** + +| What doesn't change | What changes | +|---------------------|-------------| +| Verb computation: skill-agnostic (same verbs appear) | Resolution layer: skill score modifies success rate | +| VerbPriorityProfile: career-determined, not skill-determined | Advanced verb exceptions: some verbs gated by minimum skill | +| Content authoring: one verb pool per context | Outcome authoring: branching based on skill-modified results | + +The "mostly C" caveat — some advanced verbs may be gated — is the one area that needs a bounded spec before implementation. If "advanced verb gating" is left unspecified, individual verb authors will make ad hoc decisions and we'll have a mix of outcome-only and availability-gated verbs with no consistent player mental model. + +**Flag:** Write the advanced-verb gating spec in Sprint 25 or 26, before individual verb specs are authored. Proposed rule: a verb is gated only if using it without the required skill would produce a nonsensical interaction (e.g., "Hack" with zero hacking skill is just failing at something you don't understand — the failure itself is incoherent as gameplay). Most social, economic, and movement verbs should stay ungated. + +--- + +### Decision 6 — Voice: culture-driven, job modifies + +**The inversion:** This decision inverts the assumed architecture. Culture is the primary voice register; job is the modifier layer. A Krenn tycoon sounds like a Krenn person who runs businesses. + +**What this means mechanically:** + +The voice card hierarchy is: +``` +Culture (base register) + └── Job modifier (adds professional vocabulary and priority concerns) + └── Situation triggers (fires the appropriate line) +``` + +This is clean. But it creates a tension with Decision 2. + +**The tension:** Decision 2 defers culture selection from character creation. Skills + bookmark only. If culture is the primary voice register but the player hasn't selected a culture, what base register does the voice card use? + +**Resolution options:** + +| Option | Mechanic | Tradeoff | +|--------|----------|----------| +| A — Bookmark implies culture | Tycoon bookmark defaults to a cultural context (e.g., Burnelli economic culture) | Simple, but cultural identity is defined by career, not creation | +| B — Culturally neutral Phase 1 voice | The voice card runs without culture modifier until culture is added | Voice feels generic in v0.2 | +| C — World assigns culture at generation | The apartment generation places the player in a cultural context; voice card reads that | Elegant but requires generator to have cultural zones in v0.2 | +| D — Single culture in v0.2 scope | Only one cultural context is populated in v0.2; the culture/job distinction is architectural but effectively single-valued | Simplest, works for v0.2, leaves the architecture right for v0.3 | + +Option D is probably the right v0.2 answer. Mellanie writes one culture-base voice card + tycoon modifier. The architecture is culture-primary, but with only one culture active, the distinction is invisible. When culture selection is added later, the system expands naturally. + +**This needs confirmation.** See Questions for Jeroen below. + +--- + +### Decision 7 — ALL NPCs generated, no named characters + +Kael doesn't exist. The smuggler's FRIEND is now a role template. This decision has cascading content architecture implications but the mechanical implications are actually simpler than I expected. + +**What this means mechanically:** + +| Before | After | +|--------|-------| +| FRIEND = Kael Davan (hand-placed) | FRIEND = role assigned by generator based on position + proximity | +| Monologue references Kael by name | Monologue templates with role-reference slots (`[FRIEND.name]`, `[FRIEND.role]`) | +| Phase Zero authored for a specific person | Phase Zero authored for a role archetype | +| Consequence engine tracks Kael-specific states | Consequence engine tracks role-states | + +For the VerbPriorityProfile: no change. Verbs are context-aware, not NPC-aware. The fact that an NPC is generated vs hand-authored is invisible to the verb system. + +For the relationship system: the generator assigns a generated NPC to the FRIEND-figure role. The relationship system tracks warmth accumulation with that NPC-ID. The FRIEND pattern (warmth accumulation → Phase Zero completion → arc gates unlock) works identically whether the NPC is Kael or [Generated-ID-4729]. + +**Interesting emergent design question:** What determines which generated NPC gets assigned to the FRIEND-figure role? Presumably: proximity to the player's starting location, appropriate social position for the career (a tycoon's FRIEND-figure might be a key supplier or a fellow small-business owner in the district), and some relationship-formation opportunity (overlapping schedules, shared location). This is a generation rule I don't see specified anywhere. It belongs in the zone identity spec Miri is building. + +--- + +### Decision 8 — Generative AI for NPC content templating + +Opens the content pipeline to AI-assisted generation. Culture vectors + tone + accents as templating dimensions. + +**What this means mechanically:** + +The content pipeline for NPC dialogue becomes: +``` +Template (FRIEND-figure warmth, line pool type, trigger context) + → AI parametrization (culture vector, tone, accent prompts) + → Generated line pool + → Trigger system (fires appropriate lines on game events) +``` + +This is different from Mellanie's current pipeline, which authors individual lines directly. The template layer is new infrastructure. + +**For v0.2:** "Limited vocabulary acceptable at first" means we can start with a small template set and a small generated pool. The architecture should support expansion; the initial scope should be deliberately narrow. + +**Implementation note:** The AI generation step doesn't have to happen at runtime in v0.2. Generate the line pool offline (using the template + AI), bake the results into the game data, ship them. In-game ollama (Decision 9) is for future runtime generation. The pipeline distinction matters for sprint scoping. + +--- + +### Decision 9 — Possible in-game ollama (deferred but open) + +No design work required in v0.2. But note the architectural implication: + +If in-game ollama is ever implemented, the NPC dialogue system needs to be able to call it asynchronously without blocking the main game loop. When the time comes, this is a Tyre/server concern: dialogue requests need to be non-blocking, results cached, and the simulation tick should not wait on LLM inference. + +Flag this as an architectural constraint to keep in mind, not a v0.2 implementation item. + +--- + +### Decision 10 — Quietly responsive world, not indifferent + +Gore's Kenshi-indifference premise rejected. The world notices locally. Gradient: world → district → neighbors → colleagues → friends. + +**What this means mechanically:** + +This isn't just a content decision — it's a simulation behavior decision. The "quiet responsiveness" Jeroen describes (prices shift when you buy, NPCs mention you were there yesterday) isn't authored content arriving in Phase 2. It's Phase 1 simulation behavior. + +**Systems that produce quiet responsiveness:** + +| System | Behavior | Complexity | +|--------|----------|------------| +| Economy tick with local price elasticity | Buying from a vendor shifts that vendor's prices for a period | Low — add price elasticity factor to economy tick | +| NPC short-term memory | NPCs log recent player encounters; surface this in greetings after threshold | Medium — NPC state component, decay timer | +| Relationship proximity accumulation | Repeated location sharing accumulates a "familiarity" score | Low — already part of relationship system | +| District reputation | Local actions visible to district NPCs aggregate into a district-level known status | Medium — reputation component, district-scoped | + +These belong in Phase 1 server work, not Phase 2 authored content. The Phase 1 world needs these simulation behaviors to produce the "latent responsiveness" Gore correctly identified as necessary before Phase 2 consequence weight can land. + +**The gradient design:** + +``` +World (global) → Faction reputation (very slow, very public actions only) +District (local) → District reputation (consistent presence, notable actions) +Neighbors/colleagues → Relationship warmth (regular interaction) +FRIEND-figure → Deep relationship state (Phase Zero accumulation) +``` + +This gradient is the map of what the consequence engine tracks. The storyteller reads this gradient when deciding whether to inject pressure. + +--- + +### Decision 11 — Full character customization + +Hair, clothing, colors. Readability through outline/highlight, not by limiting customization. + +**What this means mechanically:** + +The character creator needs appearance variables: hair type, clothing type, color palettes. These are distinct from the NPC archetype system. NPCs are legible through archetype-silhouette + behavioral state (Araminta's system). The player character is legible through their custom appearance + career insert visual grammar. + +No systems change needed beyond the character data model gaining appearance fields. The tile renderer already handles visual output. The outline/highlight solution is a visual design choice, not a systems architecture choice. + +--- + +### Decision 12 — Setting delivery: both layers (visual + insert) + +Visual shows it; insert names and contextualizes. Parallel production tracks. + +**What this means mechanically:** + +The insert becomes the primary setting delivery mechanism for information the physical world can't show. Zone type, NPC names, prices, your own economic state — these come through the insert, not through ambient world reading. + +For systems design, the insert is a designed UI system that reads from the observer snapshot. The insert's information content is career-filtered: a tycoon insert shows economic data, a law enforcement insert shows procedural data. The snapshot must carry enough world-state for the insert to populate correctly. + +This is already part of the diegetic tool suite design. The tycoon insert is the first one we need to spec for v0.2. + +--- + +### Decision 13 — First Settled Reach moment: apartment + insert activation + +Two moments. Before you leave the room. + +**What this means mechanically:** + +Two scripted events, fired in sequence on first session: + +1. **Apartment generation:** The world generator places the player in housing appropriate to their starting economic state. For tycoon: moderate standing (not wealthy, not destitute — you're building, not arrived). The apartment READS as your economic position through visual density, furnishing quality, view of the district. + +2. **Insert activation:** A designed UI event. The insert powers on, career UI appears for the first time. The tycoon insert activates with its economic dashboard. This is simultaneously: setting delivery (the insert contextualizes where you are), career identification (your economic tools are real, you're a tycoon), and tutorial (through use, not instruction). + +**Design question:** What drives the apartment's economic class in character creation if we only have skills + bookmark? The tycoon bookmark probably defaults to a moderate economic starting state. But can the player influence this through skill allocation? A character with high social manipulation might start with better social connections (smaller apartment but better district) vs high hacking (better tools, worse physical space). This is flavor if skills are outcome-only (Decision 5 says mostly C), but it's the kind of detail that makes character creation feel like it matters. + +--- + +### Decision 14 — Groundhog Day alarm clock homage + +First day only. *Click* pa-pa pa-pa, cut short. "New day, new start, new chances." + +**What this means mechanically:** + +One-shot event: `first_session_day_1 = true`. The alarm fires on Day 1, game boot 1. After that, it's a standard day-start notification (or silence — the routine has absorbed it). + +This is the authored beat that opens Phase Zero. Miri notes correctly that the `day_start` trigger type is needed for the monologue system. The very first monologue line should fire here — establishing the character's voice before any world interaction has occurred. + +The Groundhog Day structure (alarm → calendar ping → appointment) is the authored scaffolding that teaches the tycoon bookmark through the first day's rhythm. This is the "rails to take off from" (Decision 15) — the authored first beat, then agency. + +--- + +### Decision 15 — Player choices ARE the content (Rimworld model) + +This is the philosophical foundation that changes what several other systems need to do. + +**What this means mechanically:** + +There is no formal mission system. There is no quest tracker. There is no objective list. The job (tycoon) provides a context (I'm building a business) and access to tools (economic verbs, insert dashboard). Everything that happens is a consequence of player choices within that context. + +The consequence engine (CauseChain tracking) becomes the primary authored system. The storyteller's job is pressure calibration: when to inject authored pressure (the ring needs a new supplier — the tycoon is approached), when to let quiet days run. + +**What "a job is rails to take off from" means architecturally:** + +| What the job provides | What it doesn't provide | +|----------------------|------------------------| +| VerbPriorityProfile (economic verbs salient) | Objectives or tasks to complete | +| Diegetic tools (insert dashboard, investment terminal) | Quest markers or waypoints | +| Starting location and FRIEND-figure context | A pre-authored narrative thread | +| Income tick and economic foothold | A tutorial beyond the first day | +| Asymmetric lens (what the tycoon sees vs others) | The story — the player makes the story | + +The storyteller reads the simulation state and injects opportunities: an NPC approaches with a deal, a district event creates an economic opening, a relationship threshold triggers a new interaction possibility. The player decides what to do with those opportunities. + +**The consequence engine scope question:** If every player choice is potentially content, what does the CauseChain track? In a mission-based game, this is clear (mission decisions are trackable events). In the Rimworld model, the CauseChain needs to recognize which player actions are "consequential" — worth tracking for future reference. I don't have a spec for this. See Questions for Jeroen. + +--- + +## Part 2: Tensions and Contradictions + +### Tension 1: Culture-driven voice + deferred culture selection + +**Decisions in tension:** Decision 6 (culture-driven voice) + Decision 2 (culture deferred from character creation). + +**The problem:** Voice is culture-primary. But the player doesn't select culture in v0.2. The voice card has a base register it can't derive. + +**My recommendation:** Decision D (single cultural context in v0.2, architecture is correct but single-valued). Design the voice system to accept a culture parameter. For v0.2, hardcode one value. When culture selection arrives in v0.3+, the system already supports it. This costs nothing architecturally and costs Mellanie one voice card instead of many. + +**But this needs explicit confirmation** — because if Jeroen means the tycoon bookmark IS a cultural statement (tycoons in this world are Burnelli-economic in register by definition), then culture comes through the bookmark and the tension dissolves differently. + +--- + +### Tension 2: Generator-first proof-of-life vs earliest-possible playtest + +**Decisions in tension:** Decision 1 (proof-of-life = generator + graphics) vs the v0.1 lesson (we built too much before testing the fun hypothesis). + +**The problem:** Tyre's Tier A (hand-built world, 2-3 sprints, first playtest by sprint 28) is rejected. Tier B (template-generated) is the minimum. That means the first playable proof is later. + +**I am not recommending against Decision 1** — the generator IS the vision and Jeroen is right that testing on a hand-built world proves nothing about the generator. But I want to flag: the team should have a planned "first playable state" moment in the generator-first roadmap. What does the first thing you can actually play look like, and when does it arrive? If that's sprint 30, that's fine — but it should be planned, not discovered. + +--- + +### Tension 3: "Quietly responsive" requires Phase 1 simulation work, not just authored content + +**Decisions in tension:** Decision 10 (quietly responsive world) + Decision 15 (player choices are the content, Rimworld model). + +**The problem:** If the Phase 1 world is designed to be quietly responsive (prices shift, NPCs remember), that responsiveness must be built into the simulation — not delivered through Phase 2 authored content injection. The Rimworld model's storyteller handles dramatic pressure injection, not ambient world responsiveness. These are two different systems: + +- **Storyteller:** When does authored content arrive? How intense is it? (Phase 2) +- **Ambient responsiveness:** The simulation's baseline consequence behavior. (Phase 1) + +If we assume "quietly responsive" is delivered by the storyteller's authored content, we'll get Phase 1 as a tech demo that feels unresponsive until Phase 2 content arrives. Gore correctly identified this as the emotional register risk. The fix is simulation-level, not content-level. + +**Action required:** The Phase 1 server work plan needs to include the ambient responsiveness behaviors (price elasticity, NPC memory, familiarity accumulation) as simulation features, not as authored content placeholders. + +--- + +### Tension 4: "Mostly C" skills + "some verbs gated" = unspecified exception surface + +**Decision in tension:** Decision 5 ("mostly C" with unspecified gated exceptions). + +**The problem:** "Mostly C" with exceptions that need a spec is a design debt item. It sounds minor but it's actually a correctness invariant: the player's mental model of "I can always try anything" is violated every time they encounter a gated verb without understanding why. Inconsistent gating is more confusing than either full gating or no gating. + +**Action required:** Before any verb specs are authored, define the gating rule. My proposed rule: a verb is gated only when the action is incomprehensible without the required capability (not just ineffective — the action doesn't make sense). "Hack a terminal with zero hacking skill" = nonsensical, gate it. "Negotiate poorly with low social skill" = meaningful failure, don't gate it. + +--- + +## Part 3: Implementation Risks + +### Risk 1: Tycoon verb design is a blank page + +**The highest priority new design work for v0.2.** We have zero designed tycoon verbs, zero tycoon VerbPriorityProfile, zero tycoon economic interactions. All previous verb work (smuggler, law enforcement) is irrelevant to the v0.2 scope. This needs to be the first concrete design session. + +**Who needs this:** Tyre (verb computation), Araminta (verb prompt visual treatment), Mellanie (trigger catalog tied to verb outcomes), Ozzie (first day design needs tycoon-specific tools). + +--- + +### Risk 2: NPC generation rules are upstream of everything + +The generator needs to know what NPC archetypes exist and what roles they fill in each zone type. This isn't Miri's zone identity spec alone — it's the intersection of zone identity + NPC archetype design + role assignment rules. If this spec doesn't exist before the generator runs, the generator produces either no NPCs or random NPCs that don't fit the world. + +**The spec needs:** +- Zone types and their NPC archetype distributions +- Which archetypes can fill which roles (FRIEND-figure, competitor, supplier, customer...) +- What physical and behavioral characteristics mark each archetype +- The assignment algorithm for FRIEND-figure role selection + +This is a joint spec: Miri (zone identity + cultural meaning), Gestalt (role taxonomy + assignment rules), Araminta (behavioral and visual markers per archetype). One document, three contributors. + +--- + +### Risk 3: The consequence engine scope for a mission-less game + +Without a formal mission system, the CauseChain system has no obvious boundary for what it tracks. In the Rimworld model, every player action is potentially a seed for future consequences. But the CauseChain system can't log everything — it needs a selective definition of "consequential action." + +**Proposed initial scope (to be confirmed):** +- Economic transactions above a threshold (large investments, major deals) +- NPC relationship-state changes (first meeting, relationship threshold crossings, conflict events) +- World-state changes triggered by player actions (taking or losing a business asset, district reputation changes) +- Storyteller-injected events (authored pressure that the player responds to) + +This is a fraction of all player actions, but it captures the events that are likely to produce future consequence. The risk is under-tracking (important seeds are missed) or over-tracking (performance and signal-noise problems). + +--- + +### Risk 4: The apartment generation requires economic class data at character creation + +Decision 13 says the apartment reflects the player's economic position. Decision 2 says character creation is skills + bookmark only. What economic starting state does the tycoon bookmark imply? + +If the tycoon bookmark has a fixed economic starting state, the apartment generation is straightforward. If skills can influence starting economic state (high social = better connections, better starting district), the apartment generation becomes a function of character creation parameters. + +This needs to be specified before the apartment generator is built. It's a small decision but it affects both the character creator (what does skill allocation change?) and the world generator (what economic class input does the apartment generation take?). + +--- + +## Part 4: Questions for Jeroen + +### QUESTION FOR JEROEN 1: How does the tycoon bookmark engage with culture in v0.2? + +**The tension:** Decision 6 says voice is culture-driven. Decision 2 defers culture from character creation. The voice system needs a base culture register to operate. + +**The concrete question:** For v0.2 with the tycoon bookmark as the only career: + +- Does the tycoon bookmark *imply* a cultural register (tycoons in this world are economically Burnelli-adjacent by definition, so picking tycoon picks your voice base)? +- Or is the v0.2 culture layer simply "one culture, not selectable," with the culture/job distinction existing architecturally but invisible until culture selection is added? +- Or is there a third option — the world generator places you in a culturally specific starting zone (culture comes from the world, not creation), and the voice card reads the zone's cultural context? + +**Why it matters now:** Mellanie can't write voice cards until she knows whether to write one culture-base + tycoon modifier, or to write a parametric template, or to write the tycoon voice as culturally neutral until culture is added. The answer changes her production scope significantly. + +--- + +### QUESTION FOR JEROEN 2: What does a tycoon DO at the verb level? + +**This is the most urgent design gap.** We've confirmed tycoon as the v0.2 bookmark. But we have zero designed tycoon verbs, zero VerbPriorityProfile, zero economic interactions specified. + +**The concrete question:** When a tycoon player approaches an NPC in the world, what is the default action offered? What economic interactions look like at the verb level? Is "Negotiate" a single verb that encompasses deal-making, or is there a vocabulary of economic verbs (Invest, Hire, Sell, Buy, Inspect, Verify)? + +And for the tycoon's world footprint: who are the NPC types the tycoon primarily interacts with? Suppliers, customers, competitors, district officials, financiers? The VerbPriorityProfile needs to know what the tycoon's world looks like from a verb interaction standpoint. + +**Why it matters now:** The tycoon VerbPriorityProfile is the first design deliverable needed before: Tyre can spec the verb computation, Mellanie can write trigger-catalog content, Ozzie can design the "First Day" tycoon experience, and Araminta can design the insert visual grammar. Everything downstream waits on this. + +--- + +### QUESTION FOR JEROEN 3: What does the CauseChain system track in a mission-less game? + +**The tension:** Decision 15 (player choices ARE the content, Rimworld model) removes the formal mission system. Without mission objectives as trackable events, the CauseChain system has no obvious boundary for what constitutes a "choice worth tracking." + +**The concrete question:** When the consequence engine remembers "this happened because of your decision," what categories of player decision are being tracked? Is it: + +- **Economic thresholds** (you invested significantly in X, so when X is threatened it's your problem) +- **Relationship events** (you helped/harmed NPC Y, so Y now has a stance on you) +- **World-state changes** (you triggered action Z which changed district state, so Z's downstream effects are attributed to you) +- **All of the above** (and the storyteller filters for what's narratively interesting) + +The consequence engine's performance and signal-noise ratio depends on how selective this tracking is. In Rimworld, the storyteller has access to all simulation state — it doesn't log "decisions," it just reads current state. If we're modeling consequences as stored cause chains, we need a policy for what gets recorded. + +**Why it matters:** The answer determines whether CauseChain is a lightweight event log (high performance, potentially missing causes) or a comprehensive player history (richer consequences, more memory). For a Rimworld-model game without mission checkpoints, I'd lean toward reading current world state rather than logging past decisions — but that's a design hypothesis, not a decision. + +--- + +## Part 5: Reactions to Other Agents' Round 4 + +### On Tyre's estimate revision + +Tyre revised to 6-10 sprints and proposed the Tier A proof-of-life sprint before the full buildout. Decision 1 rejects Tier A. This doesn't invalidate Tyre's instinct — the concern about building too much before testing is correct. But the answer to that concern is: what does the first playable state look like within the generator-first approach? Not "skip the generator for a sprint." The generator needs to reach a "bare minimum playable" state earlier than its "production quality" state. That's the milestone to plan toward. + +### On Ozzie's "First Day" beat sheet + +Ozzie's 7-beat first 30 minutes (character creation → alarm → appointment → insert activation → first work moment → first anomaly → first consequence seed) is the right structure for the tycoon bookmark. The tycoon's first work moment is probably: arriving at a business location, making your first economic decision (take the deal? inspect the asset? negotiate the terms?). The consequence seed is embedded in that first decision — a deal made today that will echo later. This is a design session that Ozzie, Paula, Mellanie, and I need to run together. I agree with Ozzie: one beat sheet, four domains. + +### On Mellanie's three monologue cadences + +Mellanie correctly flags that the tycoon bookmark (Active + WFH + Gig in one career) produces three different monologue rhythms. For v0.2, I'd recommend designing the trigger catalog for the most common tycoon rhythm first (probably Gig — accepting deals is the episodic unit of tycoon play) and treating Active and WFH triggers as secondary content. The architecture should support all three; the first content pass can prioritize one. + +### On Gore's Phase 1 emotional register warning + +"Don't let Phase 1 become the game's default emotional register." This is the most important design note in Round 4 from a systems standpoint. The Phase 1 quiet responsiveness (prices shift, NPCs remember) is the mechanism that prevents Phase 1 from feeling like a tech demo. Gore is right that if the player settles into "this world runs without me," Phase 2's authored pressure feels like an intrusion. The ambient responsiveness simulation features are the answer — not authored content, simulation behavior. + +### On Nigel's career-aware content distribution question + +Nigel's Rimworld-model question (does the storyteller seed content career-aware, or does career determine the player's lens on universal content?) is elegant and I think the answer is: both, by design tier. + +**Phase 1:** Content is distributed universally. The world has economic threads, social threads, institutional threads all running simultaneously. Career determines the player's *lens* on those threads — what's visible to them. + +**Phase 2:** Authored pressure can be career-aware when it needs to be. A tycoon gets approached about the deal; a law enforcement player gets assigned the investigation of the same deal. Same authored ingredient, career-specific surface. + +This is the right architecture because it doesn't require separate authored content per career — it requires career-filtered visibility on shared world content. That's cheaper to produce and more emergent in behavior. + +### On Miri's zone identity spec as NPC population spec + +Miri correctly identifies the zone identity spec as the upstream dependency for the generator. I want to extend her point: the zone identity spec isn't just a worldbuilding document. It's also an NPC archetype distribution spec. For the generator to produce a working-class logistics zone, it needs to know: what archetype NPCs fill that zone, in what proportions, and what roles they fill. The FRIEND-figure assignment rule (which NPC in the tycoon's starting zone becomes their FRIEND) is part of this spec. + +Miri's spec should have three sections: (1) visual/physical zone identity, (2) NPC archetype distribution per zone type, (3) economic behavior rules per zone type. This is a cross-domain document. I'd like to co-author it. + +--- + +## My Single Most Important Recommendation + +**Design the tycoon verb map and VerbPriorityProfile before any other v0.2 design work begins.** + +Every other domain is waiting on this: +- Tyre needs the verb taxonomy to spec the computation +- Mellanie needs the VerbPriorityProfile to write trigger-catalog content +- Araminta needs the tycoon verb prompt visual treatment +- Ozzie needs the tycoon tools to design the First Day +- Paula needs the tycoon FRIEND-figure context to spec Phase Zero +- I need it to confirm whether skills (Decision 5's "mostly C") need any gated exceptions in the economic verb space + +The tycoon bookmark decision (Decision 4) resolved which bookmark. But "tycoon" is a concept, not a spec. A detective has "Investigate, Observe, Interrogate." A tycoon has... what? That question needs to be answered before Sprint 25 work planning, not during it. + +This is the decision that costs a day to make and weeks to untangle if made wrong mid-implementation. + +--- + +*GESTALT — Round 5 complete. The 15 decisions are sound. The tycoon pivot is bold and right. The generator-first approach is ambitious and correct. The systems tensions I've flagged (culture/voice, Phase 1 responsiveness as simulation behavior, CauseChain scope without missions) are solvable — they just need explicit decisions. Get Jeroen's answers on the three questions above and I can finalize the tycoon VerbPriorityProfile spec within the same sprint.* diff --git a/docs/workshops/wheres-the-fun/round5-gore.md b/docs/workshops/wheres-the-fun/round5-gore.md new file mode 100644 index 000000000..be9d27050 --- /dev/null +++ b/docs/workshops/wheres-the-fun/round5-gore.md @@ -0,0 +1,363 @@ +# Round 5 — Gore: Reacting to the 15 Locked Decisions +## Where's the Fun? Workshop | 2026-03-05 + +**Agent:** Gore (Themes & Endgame Design) +**Round:** 5 — Decision Reactions + +--- + +## Orientation: What Round 4 Resolved and What It Opened + +My Round 4 carried three questions into the interview: + +1. Kenshi-weight vs. social-weight — which governs consequence? +2. Is the Phase 1/Phase 2 seam designed as a threshold or invisible? +3. Does the transhumanist ladder require a visible human skill ceiling? + +Decision 10 (quietly responsive) answers question 1 directly — and dissolves question 2 in a way that changes my entire Phase 1/Phase 2 framing. Decision 4 (tycoon is the v0.2 bookmark) answers none of them, but opens a new one I hadn't asked: what does the moral arc look like for a career without personal complicity at its center? And Decision 2 (skills + bookmark only) defers the transhumanist ladder, but doesn't answer question 3 — the architectural decision needs to be made now even if the content waits. + +I'll go through all 15 decisions. Then I'll do three deep dives on the most consequential open threads. Then questions for Jeroen. + +--- + +## Decision-by-Decision Reactions + +### Decision 1: Proof-of-life = generator + graphics + +The generator is not an engineering milestone. It is the thesis statement. + +Every science fiction world asks its audience to accept a world that runs independently of them. Most games fake this with authored set-dressing. The generator actually does it — locations emerge from geographic characteristics, NPCs fill roles because roles need filling, the economy ticks for internal reasons. When the generator works, the player is not in a stage set. They are in a world. + +This matters for themes because consequence only has weight in a world with independent existence. If the world was made for you, your choices were expected. If the world runs without you, your choices are intrusions the world has to accommodate. The generator is the thematic prerequisite for everything else. + +**One flag:** The generator needs to produce the Settled Reach's specific texture, not generic sci-fi urbanism. Post-scarcity abundance, species and cultural variety, the particular mix of mundane contentment and latent ambition — these aren't emergent from a working life-sim engine. They require zone identity work (Miri's domain) to be baked into the generation inputs. A generator running without zone identity proves the technology. A generator running with zone identity proves the world. + +Zone identity is not optional for the proof-of-life milestone, even though it looks like it is. + +--- + +### Decision 2: Skills + bookmark only. Decision 3: Religion not a system. + +Correct scope for v0.2. But there's a separation I want to flag before it becomes a confusion. + +Culture is deferred from *character creation*. It is not deferred from *NPC generation*. The generator needs cultural vectors to produce characters who feel like they belong to specific peoples and places. Miri's world culture taxonomy is a dependency for the generator even though the player doesn't select a culture at character creation. These are different deferrals pointing in different directions. Conflating them would leave the generator producing culturally undifferentiated NPCs because we declared culture "deferred." + +On religion: correct. Religion in the Commonwealth universe is one of many answers to "what is intelligence for?" — but in a game about personal ascension rather than civilizational narrative, institutionalized religion is too diffuse to carry weight. The game's question ("what is intelligence for?") is still answered through play, through career choice, through the transhumanist ladder. It doesn't need a church. It IS the church. + +--- + +### Decision 4: Tycoon is v0.2 bookmark, zero investigation + +The most important thematic decision in the 15, because it changes the moral arc's structure entirely — and the existing arc architecture was built for the smuggler. + +The tycoon is actually a *richer* vehicle for consequence-as-theme than the detective or smuggler, but it requires a different complicity model. + +The smuggler's complicity is **personal**: you are doing something morally questionable, and the weight accumulates as you understand its human cost to a specific person you know. Kael is real to you before Naia's situation makes his vulnerability legible. + +The tycoon's complicity is **systemic**: you are building something, and the building participates in systems that harm people you don't necessarily see. You aren't choosing to hurt anyone. You discover that you were already hurting people through the aggregate effects of your choices. This is how complicity actually works in the world — not theatrical villainy but supply chains, investors, business partners, all connected before you understood the connections. + +The tycoon's three career modes (Active/WFH/Gig) map naturally onto three modes of moral exposure: +- **Active** work puts you in direct contact with the consequences of your decisions — the employee at their desk, the product leaving your hands +- **WFH/remote** means you're making decisions without seeing their effects — capital deployment at a distance +- **Gig** means taking contracts without necessarily knowing their full context + +Three modes of doing things whose consequences are variably visible. That's rich thematic material. + +**The structural problem:** The entire moral arc architecture built over three workshop rounds is designed around the smuggler's Phase Zero — warmth with Kael as a specific named person, the crack arriving through Naia's situation, the reckoning triggered by your understanding of their entanglement. That architecture requires a *named FRIEND* whose wellbeing is specifically entangled with your choices. + +For the tycoon, there is no Kael. There is no hand-authored FRIEND. All NPCs are generated (Decision 7). The tycoon's Phase Zero equivalent needs to be designed from scratch, and the first design question is: **who is the tycoon's Kael?** + +This is not a rhetorical question. See the Questions for Jeroen section. This is the first design question that needs an answer before Paula's team can author the tycoon's Phase Zero, before Ozzie can design the tycoon's Consequence moment, before Mellanie can write the tycoon's monologue arc. + +--- + +### Decision 5: Skills affect outcome (mostly C) + +Thematic endorsement, plus a note about where the transhumanist hook lives within this decision. + +"Everyone sees the same verbs. Skills determine how well you do." This is the post-scarcity design ethos made mechanical. The world doesn't gate you. What you bring to a situation determines what you get from it. A poor negotiator who tries to negotiate and fails badly creates more interesting consequence than a locked verb — the failure is what generates the story. + +There's a subtler implication: if skills affect *quality* rather than *access*, skill investment becomes a statement about character identity as well as tactical preference. The character good at social manipulation handles situations one way. The character good at mechanical repair handles them another. Your competencies shape the kinds of consequences you produce — which connects consequence back to character. + +**The transhumanist note within this decision:** When you go Higher, you gain perception modes that are literally impossible for baseline humans (D-017). This isn't a skill improvement — it's a different category of capability that doesn't exist in the baseline verb list. The system should have two distinct tiers: +- **Human verbs:** modulated by skill +- **Post-human verbs:** unlocked by qualitative character transformation + +These should not look like the same system. The first post-human verb the player gains should feel like stepping through a door, not like leveling up a skill. + +--- + +### Decision 6: Culture-driven voice, job modifies + +Strongly endorsed. + +Culture is the answer your civilization converged on to "what is intelligence for?" A Krenn person's voice IS a particular answer to that question. Their job expresses it through a specific domain. You are your background before you are your career. The inversion from Mellanie's original proposal (job base, culture modifier) to Jeroen's (culture base, job modifier) corrects a category error: jobs change, background doesn't. + +One thematic note for v0.3+ planning: if a character goes Higher, does their cultural voice persist? Does neural modification change who they sound like, or does the culture survive the upgrade? The transhumanist ladder is thematically most powerful if cultural voice is the last thing to change — or the thing that finally changes, signaling that the character has become something other than who they were. The moment the voice stops sounding like a Krenn person who runs businesses is the moment the player knows the transformation is complete. + +Flag for later. Don't implement now. But design with it in mind. + +--- + +### Decision 7: ALL NPCs generated. No named characters. + +The most consequential thematic decision in the 15. I want to go deep on it — see Deep Dive 2 below. Short reaction here. + +Kael should not exist. Jeroen is right. The generator should produce an NPC who fills the smuggling-operation's logistics manager role because the location's geography makes smuggling viable — and that NPC then accumulates a specific identity through the data the generator assigns them (partner, specific stress points, particular behavioral tells) and through the player's observation of them. + +I support this. But I want to be precise about what the generator needs to produce: not just attachment (Rimworld does attachment), but what I'm calling *moral particularity*. The distinction matters for whether the Phase 2 crack can carry the weight the design needs it to carry. + +--- + +### Decision 8: Generative AI for NPC content templating + +The cultural template vector approach is right. But the thematic bar for what the templates need to accomplish is higher than the decision text implies. + +The templates need to encode **worldview**, not just register. What does it mean to be from a particular culture in the Settled Reach? Not just accent and tone — but what you notice, what you leave unsaid, what you consider important enough to mention, what silences mean. Cultural vectors that encode worldview produce characters who feel *locally rooted*. Cultural vectors that encode only register produce characters who feel locally *accented* — distinct but shallow. + +A second axis the templates need: **NPC stakes**. What does this person have to lose? What do they want? What are they afraid of? These situational dimensions are what separate a culturally distinct NPC from a morally particular one. The generator doesn't need to know Kael's name — it needs to know that the logistics manager has a partner in a fragile situation, a principled refusal about certain cargo types, and a relationship with the player's character that's accumulated three weeks of small trust signals. These stakes are what make the Phase 2 crack matter. + +--- + +### Decision 9: Possible in-game ollama for live NPC dialogue + +The most thematically transformative and most dangerous option in the feature set. + +Transformative: an NPC who responds to your specific choices with language specific to their generated personality produces genuine interpersonal entanglement. If the logistics manager can be surprised by your betrayal, confused by your inconsistency, hurt by what you said last week — the moral weight problem becomes much easier to solve. The player can't pretend the NPC is a prop if the NPC is demonstrably a continuous, responsive entity. + +Dangerous: the same system that generates responsiveness also hallucinates. An NPC who breaks character at a morally critical moment shatters the weight. You cannot carry weight about a character who stops being coherent. Moral weight requires narrative consistency, and LLMs are not reliably consistent. + +My design constraint for when this ships: ollama should operate *within* authored template constraints, not *instead of* them. The templates (Decision 8) define the character's range — their vocabulary, their stakes, their cultural register, their emotional limits. Ollama provides constrained improvisation within that range. Not freeform dialogue; templated responsiveness. This is how you get dynamic responsiveness without character-consistency collapse. + +Design for the transition now. The knowledge graph architecture should be readable by a future LLM dialogue system. Don't build walls between the current and future states. + +--- + +### Decision 10: Quietly responsive, not indifferent + +This decision answers my Round 4 Q1 and dissolves my Round 4 Q2 in one move. + +In Round 4, I proposed a Phase 1/Phase 2 split: Phase 1 is Kenshi-weight (world indifferent, consequences are internal), Phase 2 is social-weight (world responds, consequences are external). My concern was the seam between them — a player who has learned indifference as the rule will experience the first authored responsiveness as an intrusion. + +Decision 10 dissolves the seam. The world is **always** quietly responsive — at a gradient based on social proximity: + +``` +World (barely notices) → District (registers your pattern) → Neighbors/Colleagues (track you) → Friends (actively care) +``` + +This isn't a phase transition. It's a standing architecture. The authored content in what I was calling Phase 2 doesn't introduce responsiveness — it amplifies responsiveness that was already present at the local scale. The player has been living in a quietly responsive world from day one. The moral arc escalation doesn't contradict their learned model; it deepens it. + +This is a better design than my split model. It's also harder to execute, because **the gradient needs to be legible**. A world that is described as quietly responsive but experienced as indifferent is a broken promise. The player needs to observe the gradient building — to notice that someone noticed them — before the authored pressure arrives. + +See Deep Dive 1 for the legibility design challenge. + +--- + +### Decision 11: Full character customization + +Identity investment at creation is the right design choice, and Ozzie correctly names it Wow Moment Zero. The player designs a person before they play one. + +The thematic dimension: the customization creates a stake in the transhumanist question that you couldn't have otherwise. You designed this person's face, clothing, hair. Going Higher means changing what this person fundamentally *is*. ANA means this person ceases to have a physical form at all. The creation investment makes the ladder's cost real in a way that abstract capability progression doesn't. + +This is the long-arc design working correctly: invest the player in a physical identity, then ask whether they'd trade it for transcendence. Character creation at the start ensures the player has something to answer that question *with*. + +--- + +### Decision 12: Both layers (visual + insert) + +The visual layer shows the world as it is; the insert names it. This is the game's epistemological model made literal. + +The insert's thematic job is not only setting delivery. It's also the player's relationship with their technology — their human-machine integration. For the transhumanist arc, the insert is the first rung of the ladder. It provides information externally. Going Higher internalizes the capability (you don't need the insert to read the room; you perceive what the insert used to tell you). ANA means you ARE the information layer. The insert should be designed with this arc in mind — as the first step of a progression, not just a UI tool. + +--- + +### Decision 13: First moment — apartment + insert activation + +Two first moments. Both are thematically correct. + +**The apartment:** The world's opening statement about your position in it. Not "you are the protagonist" but "you are a person at a specific point in an economic distribution." Post-scarcity means everyone has enough — but "enough" looks different at different points in the distribution. You wake up already embedded in a social-economic reality before you've made a single in-world decision. That's the right opening note for a game about consequence: you are already situated, already not free. + +For the tycoon bookmark, this apartment might be comfortable or even wealthy — a statement about what you're building from. A comfortable apartment that could have been more comfortable with different choices is already a moral fact about you before day one starts. + +**The insert activation:** Neural augmentation powering on. The first step on the transhumanist ladder, experienced as unremarkable. The world treats this as normal — as normal as putting on glasses. And yet you feel it. That small awareness seeds a question: what does it mean that this is normal? This is the right thematic entry to a game that eventually asks whether you want to climb the rest of the ladder. + +Together, the two moments establish: *you are someone, in a world, with technology that is already part of who you are*. Everything that follows is elaboration of that. + +--- + +### Decision 14: Groundhog Day alarm clock homage + +"New day, new start, new chances." With a wink. + +The thematic depth of the reference: Groundhog Day is about someone who discovers, through infinite repetition, that the only way forward is to become genuinely better — not to optimize outcomes, not to game the system, but to actually *be* a better person. Our game doesn't repeat the days. The wink is that you get one of each. The alarm clock says "new day" every morning because every day genuinely is a new chance — not because anything resets, but because you carry forward who you were yesterday. + +"Click pa-pa pa-pa, cut short, first game day only" is the right constraint. This sound is a covenant between the player and the game on day one. It should never fire again. Every subsequent morning, the alarm is generic. The contrast between the first day's specific sound and every subsequent morning's generic alarm is a miniature version of the game's thematic arc: the first day is authored, every day after is yours. + +--- + +### Decision 15: Player choices ARE the content + +The fullest philosophical statement of what this game is. And one I want to be careful with. + +"Player choices are the content" can be read two ways: + +**Reading A (pure Rimworld):** The world generates conditions; the player responds; the responses are the story. Whatever happened is the narrative. This produces emergent narrative. + +**Reading B (Settled Reach):** The world generates conditions; the player responds; the choices the player makes reveal something — about what they value, about what they're willing to do, about who they're becoming. Whatever happened carries *weight*, not just texture. This produces emergent meaning. + +Reading A can be satisfied with good systems design. Reading B requires the world to respond in ways that let the player reflect on what they did — to feel the weight while they're still in it, not only when they're telling the story afterward. + +The mechanisms that produce Reading B: the journal surfacing patterns the player didn't notice consciously, the monologue commenting on something the player just did without being asked, the quietly responsive gradient showing that someone else registered the weight before you did. These turn "what happened" into "what it means." + +Rimworld players mostly read meaning *into* events from the outside. The Settled Reach's thematic ambition is to build meaning *into* the experience itself. Same world-generation model, different relationship to the player's emotional processing of it. + +--- + +## Deep Dive 1: The Quietly Responsive Gradient — Legibility and Directed Experience + +Decision 10's gradient is the emotional spine of the game. The implementation challenge is making it legible without making it instrumental. + +**The legibility problem.** The gradient needs to be visible to the player, or they can't invest in it. A player who doesn't notice that the district NPC's greeting changed — that this colleague is transitioning from "stranger" to "person who's paid attention to you" — can't make informed choices about that relationship. They might harm it accidentally. They might fail to invest when they could. + +Three signal channels the gradient can use: + +- **Behavioral tells (simulation output):** The NPC greets you by name. Mentions something you told them last shift. Adjusts their route to cross paths with you. Pure simulation — no HUD required. Highest fidelity but lowest legibility for first-run players who are still learning to read the screen. + +- **Monologue commentary (interior register):** Your character notices the behavioral change and voices it. "She remembered my name today." "He saved a seat." This is the most emotionally natural signal — it's what a person actually thinks when someone pays attention to them. Requires the monologue system to be calibrated for social signals, not just environmental or mission-state ones. + +- **Thread tracker/journal surfacing:** A relationship threshold is crossed; the knowledge graph shows that this NPC now has a "colleague — mutual recognition" tag rather than "stranger — no data." Confirms what the player has already perceived through the other two channels. Should not lead — should confirm. + +**My recommendation:** Primary signal is behavioral (the world produces it, the player perceives it). The monologue provides the interior register that names what the player just saw ("she stopped to ask about what I mentioned last shift"). The insert surfaces the relationship summary as confirmation, not disclosure. + +Never lead with the HUD signal. Let the world produce the evidence first. Then let the character name what they noticed. Then let the insert confirm it. + +**The directed experience problem.** The Rimworld model says player choices are the content — no authored goals pushed at you. But the gradient needs to become a player goal without being a mission. The player needs to discover, through play, that the world is more responsive near certain people — and to begin naturally gravitating toward those relationships. + +The tycoon bookmark onboarding should embed this without scripting it: the bank relationship manager who learns your name on day one, the supplier who shows up at your location regularly. You're not told to build a relationship with them. You're simply put in proximity, repeatedly, with someone who has capacity for gradual responsiveness. The relationship-building is available. Whether you invest in it is your choice. + +--- + +## Deep Dive 2: Generated NPCs and Complicity-Grade Emotional Weight + +Decision 7 is right. But I want to name the distinction between what the reference games achieve and what this game needs. + +**Two emotional registers of NPC attachment:** + +*Rimworld colonist attachment* — territorial, functional. "My Pawn." You feel loss when they die because they were part of your project and your survival. The loss is real, but it's closer to losing a valued piece than losing a person. The characterization vocabulary can be relatively shallow because what generates attachment is shared history and functional role, not depth of understanding. + +*Interpersonal attachment requiring moral weight* — relational, specific. "I know this person. I understand what they care about. I understand their stakes." This is what complicity requires. To feel genuinely complicit in the generated FRIEND's situation, the player needs to understand them as an agent with their own goals — not just as a role that happened to matter. + +Rimworld generates the first. The Settled Reach needs to generate the second. + +The question is whether limited vocabulary can produce moral particularity. My answer: yes, but only if the vocabulary encodes the right things. The Sims generates interpersonal attachment through trait combinations and aspiration types that produce distinct, legible personalities. Rimworld generates it through backstories and relationship tracking that makes colonists specific. The minimum viable characterization for *moral* attachment isn't more dialogue — it's more **legible stakes**. + +What does the generated NPC have to lose? What do they want? What are they afraid of? What are their limits — the thing they won't do, even if you ask? + +A Krenn logistics manager who happens to have the "has dependents" trait is not morally particular. A Krenn logistics manager whose daily behavior shows care for a partner (the insert surfaces their schedule modifications, the monologue notices they always leave on time, they reference "we" in casual conversation), who has a principled refusal about certain cargo types that surfaces if you push, who responds to your third delivery with a small tell that says they're starting to trust you — that person is morally particular. + +None of that requires authored names or authored arcs. All of it requires the generator to encode situational specificity alongside cultural register: not just *who this person is* but *what situation they're in* and *what they're trying to do within it*. + +**The NPC generation template needs a stakes axis.** Want, Constraint, Fear, Limit. These four dimensions, combined with the cultural voice templates (Decision 8) and the situational role (what position this NPC fills based on location characteristics), produce the minimum vocabulary for moral particularity. + +The player's discovery of those stakes through observation — the careful stance, the behavioral tell, the knowledge graph entry that surfaces when you talk to them twice — is what produces the transition from territorial attachment to interpersonal attachment. And interpersonal attachment is what makes the Phase 2 discovery matter rather than just register. + +--- + +## Deep Dive 3: Transhumanist Hooks, the Skill Ceiling, and What "Going Higher" Should Feel Like + +The supplement notes "transhumanist hooks via skill_ceiling concept" as the themes implication of Decision 2. This is the architectural question that most directly connects v0.2 design decisions to the endgame I proposed in Round 3. + +**Two models of the transhumanist ladder:** + +*Model A — Extension:* The ladder is built on the same architecture as the skill system. Going Higher means more skill points, raised skill caps, expanded capability ranges. You're still the same kind of being, just more so. The player who goes Higher is like a very experienced character who unlocked a prestige class. The ladder is a long power progression. + +*Model B — Qualitative break:* The ladder replaces the skill framework with something categorically different. Going Higher means you're no longer operating in the character-creation budget model. The break is designed to feel dramatic: you leave the person you built behind. The skill budget is a *human* constraint. The ladder is about ceasing to be human in the character-creation sense. + +Model A is easier to implement and produces a smooth power curve. Model B is thematically correct and produces the emotional beat that justifies the entire endgame design. + +**My recommendation:** Design toward Model B in v0.2, even though the ladder isn't built yet. + +The v0.2 skill system should have a **visible, felt ceiling** — not just a soft budget limit, but something the player can actually run into during play. A character who has invested heavily in social skills discovers a social situation where their best isn't enough. Not because the content isn't balanced, but because there is a boundary between what a very skilled baseline human can do and what would require something more. The player bumps into this boundary. They're supposed to. + +This visible ceiling does three things: + +1. **Establishes felt human constraint** — the player experiences what it means to be as good as a baseline human can be, and to still encounter limits. +2. **Creates desire before the ladder exists** — "I want to be able to do this, and I can't" is the emotional precondition for the transhumanist choice feeling meaningful when it arrives. +3. **Creates resonance at the break** — when the player crosses into Higher or ANA, they remember bumping into the ceiling. They know what they left behind. That's a loss as well as a gain. + +If the v0.2 skill system is designed as a flat budget with soft limits that most players never consciously hit, going Higher will feel like "I got better at things." If it's designed with a hard, visible ceiling the player encounters during normal play, going Higher will feel like "I became something else." + +The difference between those two is the entire thematic weight of the endgame. + +--- + +## Questions for Jeroen + +### Q1: Who is the tycoon's Kael? + +The smuggler's moral arc requires a Phase Zero warmth built with a specific person whose wellbeing becomes entangled with your choices. The tycoon bookmark is our v0.2 focus — but the tycoon's moral arc doesn't have this person designed. + +Three models of tycoon complicity, each pointing to a different FRIEND-figure: + +**A — The First Employee.** Someone who took a risk on your venture. Their stability depends on your success. When you make a decision that's good for the business and bad for them, you're in complicity territory. This is personal-scale complicity — the same structure as the smuggler arc, different context. + +**B — The Supplier/Partner.** A small operator who depends on your contracts. When you optimize by finding a cheaper source, you understand what "cheaper" cost someone who was counting on you. Economic-scale complicity — the player participates in a system, then discovers its human cost. + +**C — The Community.** The district your business is anchored in. Your success displaces the ecosystem that welcomed you. You read this as expansion; they experience it as replacement. Systemic complicity — hardest to make legible, most resonant with the Settled Reach's post-scarcity themes. + +Which model? Or something else? + +This question is blocking Paula's tycoon Phase Zero design, Ozzie's tycoon Consequence moment, and Mellanie's tycoon monologue arc. The FRIEND-figure type determines what warmth gets authored into Phase Zero, what the crack arrives through, what the player is rationalizing during Phase 1. + +--- + +### Q2: Does the quietly responsive gradient have mechanical expression in v0.2? + +Decision 10 establishes the architecture: world → district → neighbors → colleagues → friends as a gradient of caring. I endorse this fully. The design question is whether v0.2 has any mechanical expression of it, or whether it's a framework for later sprints. + +Possible mechanical expressions: +- **Relationship tracking:** NPCs accumulate interaction count with the player; their greeting behavior changes after thresholds +- **Economic ripple:** Markets adjust to the player's purchasing patterns; minor price shifts for regulars +- **Social memory:** NPCs share information about the player through their own networks; someone has heard of you before meeting you +- **Behavioral tells:** A specific NPC's idle animation or pathing changes around a person they've started to track + +If the gradient has no mechanical expression in v0.2, the "quietly responsive" decision is a concept without delivery. A player experiencing the world as indifferent — regardless of what the design spec says — will learn indifference as the rule. And then the authored content will feel like an intrusion. + +Even one small mechanic — even just NPC greeting behavior changing after ten interactions — establishes the gradient as real. What's in scope for v0.2? + +--- + +### Q3: Is the skill ceiling designed to be felt? + +The transhumanist ladder is a v0.3+ feature. But the architectural decision about the skill ceiling needs to happen in v0.2, because it determines whether going Higher will feel like transcendence or like an upgrade. + +Two options: + +**Soft ceiling:** The skill system is built as a budget with soft limits. Most players won't encounter the ceiling as a felt constraint during normal play. The ceiling exists for balance, not as a designed experience. + +**Hard, felt ceiling:** The skill system has an explicit maximum for baseline humans that players encounter during a normal playthrough. They invest in a skill, they see that investment maxing out, they discover situations where their maxed-out human capability still isn't enough. The ceiling is a designed emotional beat. + +If soft: going Higher in v0.3+ feels like "I got more skill points." +If hard: going Higher feels like "I became something else." The player remembers bumping into the ceiling and knows what they left behind. + +I've been advocating for the hard ceiling since Round 1. I'm asking directly: is this the intent, and if so, does it affect any v0.2 character-creation or skill-architecture decisions? + +--- + +## My Single Most Important Observation + +The 15 decisions converge on the same thesis: *you are a person, in a world, and what you do matters because the world is real and other people are real.* I haven't found a contradiction between them at the thematic level. That's rare for a 15-decision set. + +But there's a design challenge that none of the 15 decisions address directly: **building a generator that produces things worth caring about.** + +The generator can run. The generator can produce NPCs. The harder problem is whether the generator produces a world that invites investment — situations where the player has something to lose before Phase 2 gives them something to lose *it with*. + +The tycoon's first clean deals. The first colleague who learns their name. The first business decision that goes well. A supplier who showed up when they said they would. A district that started to feel familiar. These aren't authored content. They're the ambient texture of a world the generator needs to calibrate for: not just "a functioning economy" but "a functioning economy in which a specific player is starting to have a stake." + +The v0.2 design question isn't "does the generator run?" The question is "does the generator produce conditions in which a player naturally begins to care?" If yes, Phase 2 can threaten what the player has built. If no, Phase 2 has nothing to threaten. + +The generator is the thesis statement. The calibration of the generator is the moral architecture. + +--- + +*Gore — Round 5 complete.* diff --git a/docs/workshops/wheres-the-fun/round5-interview.md b/docs/workshops/wheres-the-fun/round5-interview.md new file mode 100644 index 000000000..510e6c714 --- /dev/null +++ b/docs/workshops/wheres-the-fun/round5-interview.md @@ -0,0 +1,102 @@ +# Round 5 Interview: Decision Reaction Refinement +## Where's the Fun? Workshop | 2026-03-05 + +**Format:** Interactive interview via AskUserQuestion, questions consolidated from 9 agents' Round 5 reactions to 15 Round 4 decisions. Questions grouped into 5 thematic clusters from Qatux's consolidated Q-WTF-027 through Q-WTF-043. + +--- + +## Group 1: Culture Architecture + +**Context:** 7 of 9 agents independently flagged the tension between Decision 2 (culture deferred from creation) and Decision 6 (culture primary for voice). Five pipelines are blocked: voice cards (Mellanie), culture profiles (Miri), NPC Blueprint culture field (Tyre), cultural visual grammar (Araminta), and systems integration (Gestalt). + +### Q-WTF-029: Is culture implicit in the starting bookmark/location, explicitly assigned, or handled differently? + +**Jeroen:** "Culture implicit in location." + +**Decision 16:** Culture is implicit in the starting location. Tycoon bookmark in Krenn System = Krenn culture. The player does not select culture at creation; it derives from where the bookmark places them. This resolves the culture-everywhere-but-nowhere tension: culture IS in the game from day one, it's just not a character creation slider. The Krenn System provides the cultural context; NPC generation uses regional culture as the primary vector. + +--- + +## Group 2: NPC Personality and Relationships + +**Context:** All 9 agents flagged NPC legibility as the universal gate. Generated NPCs must have sufficient personality surface area for emotional attachment. This group addresses what "sufficient" means. + +### Q-WTF-034/035/036: What is the minimum NPC personality surface area? How do relationships form? + +**Jeroen:** "Traits + behavior first. Friendships and relationships build Sims-style through interaction and Rimworld-style through shared adversity. Mostly the only true metric is what the player feels, but we need to codify relationships for systems to fire." + +**Decision 17:** NPC personality starts with traits and observable behavior. Relationships form through two channels: Sims-style accumulation through repeated interaction, and Rimworld-style bonding through shared adversity (surviving a crisis together, helping each other). The player's subjective feeling is the real metric, but relationships must be codified in the system so that game systems (storyteller, consequences, NPC behavior changes) can reference relationship state. + +### Q-WTF-043: Is the tycoon moral arc authored or emergent? + +**Jeroen:** "Fully emergent, but I acknowledge that this will feel artificial. The test for me will be, does the game convey this is a relationship of my character and how do I read that. The full flavor and generated content will get in the way of properly evaluating the strength of the generator. I want that to be rock solid and usable, and then we will add truly interesting threads to pull to the universe." + +**Decision 18:** Fully emergent moral arc for v0.2. No authored arc structure for the tycoon. The generator must first prove it can produce readable relationships before layered content is added on top. The explicit test: can the player tell "this is a relationship my character has" from generator output alone? Full flavor and generated content will be layered in later, but only after the relationship foundation is proven solid. This is a deliberate proof-of-concept sequence: generator proves relationships are readable -> then add narrative depth. + +--- + +## Group 3: Economic Vocabulary and Tycoon Start + +**Context:** Gestalt flagged the tycoon verb map as a critical path gap. Miri and Tyre need to know what the tycoon actually does on Day 1. The tycoon bookmark is locked but its verb inventory is blank. + +### Q-WTF-027: What does the tycoon DO at the verb level? What are the primary verbs? + +**Jeroen:** "Broad economic vocabulary. A detective will also have contracts (job/ship rental/etc) to hire, buy, inspect. True life needs verbs for all situations. Probably following the implementation speed of the interpreting systems. To buy something, ownership of an object needs to be registrable." + +**Decision 19:** Broad economic verb vocabulary, not tycoon-specific. Life verbs that serve all careers: buy, sell, hire, rent, contract, inspect, negotiate, invest. A detective uses contracts too (hiring informants, renting surveillance equipment). Implementation follows the speed of interpreting systems — each verb requires its backing system (ownership registration for buy/sell, contract tracking for hire/rent). This is a life-sim verb set, not a tycoon verb set. + +### Q-WTF-028: What does the tycoon own or invest in on Day 1? + +**Jeroen:** "For the PoC I feel existing small business owner and reframe tycoon to that (you become tycoon with growth over time, or you sell out and go exploring). A true tycoon start would feel OP." + +**Decision 20:** The tycoon bookmark starts as an existing small business owner, not a mogul. "Tycoon" is the aspiration, not the starting state. The player begins with a small operation (bar, logistics contract, storage franchise) and grows into a tycoon over time — or sells out and pivots to exploration. A true tycoon start would feel overpowered and skip the interesting growth phase. The bookmark name is aspirational; the starting position is modest. + +--- + +## Group 4: Generator Scope and Consequence Model + +**Context:** Tyre proposed Sprint 25 generator spike as the critical path. Multiple agents asked about the consequence model and failure cascades. + +### Q-WTF-037/038: Is the consequence model Rimworld-sharp or DF-slow? What constitutes structural variety? + +**Jeroen:** "Rimworld manages it." + +**Decision 21:** Both consequence models at different scales. Rimworld-style sharp events (raids, crises, dramatic reversals) AND Dwarf Fortress-style slow accumulation (gradual relationship erosion, creeping debt, reputation shifts). Sharp events create drama; slow accumulation creates texture. Rimworld already manages both — sharp storyteller events on top of slow colony degradation. The Settled Reach follows the same dual-scale model. + +### Generator spike confirmation + +**Jeroen:** "Yes, generator spike first." + +**Decision 22:** Generator spike confirmed for Sprint 25. The generator proof-of-life is the first thing built. If the generator can't produce usable output, nothing else matters. If it can, everything else has a foundation. This aligns with Decision 1 (proof-of-life = generator + graphics). + +--- + +## Group 5: Structural Variety and Skill Ceiling + +### Structural vs cosmetic variety + +**Jeroen:** "Both, at different scales." + +**Decision 23:** Both structural and cosmetic variety in the generator. Structural variety (economic landscape, faction power balance, crisis composition, NPC role distribution) operates at seed level — different playthroughs have genuinely different world structures. Cosmetic variety (NPC names, faces, apartment layouts) operates within a structure. The generator must produce both, but structural variety is the higher-priority proof. + +### Q-WTF (Gore): Skill ceiling — visible constraint or invisible until hit? + +**Jeroen:** "No ceiling in v0.2, add later." + +**Decision 24:** No skill ceiling in v0.2. Skills have no hard cap in the initial release. The transhumanist ladder (baseline -> Higher -> ANA) is a later design layer. v0.2 proves the life-sim loop without skill constraints. The ceiling concept is preserved architecturally (Tyre's `skill_ceiling` field) but not enforced in gameplay until the base game is proven. + +--- + +## Summary: All Decisions From Round 5 Interview + +| # | Decision | Resolves | +|---|----------|----------| +| 16 | Culture implicit in starting location | Q-WTF-029/030/031 — culture architecture gap | +| 17 | Traits + behavior first; relationships Sims + Rimworld style; codify for systems | Q-WTF-034/035 — NPC personality surface area | +| 18 | Fully emergent moral arc for v0.2; generator must prove relationships readable first | Q-WTF-043 — tycoon arc design | +| 19 | Broad economic verb vocabulary (life verbs, not tycoon-specific) | Q-WTF-027 — tycoon verb map | +| 20 | Small business owner start; tycoon = aspiration not starting state | Q-WTF-028 — tycoon Day 1 | +| 21 | Both Rimworld sharp events AND DF slow accumulation at different scales | Q-WTF-037 — consequence model | +| 22 | Generator spike Sprint 25 confirmed | Q-WTF-032 — generator scope | +| 23 | Both structural and cosmetic variety at different scales | Q-WTF-038 — generator variety | +| 24 | No skill ceiling in v0.2; design transhumanist ladder later | Gore skill ceiling question | diff --git a/docs/workshops/wheres-the-fun/round5-mellanie.md b/docs/workshops/wheres-the-fun/round5-mellanie.md new file mode 100644 index 000000000..c7fbc3e26 --- /dev/null +++ b/docs/workshops/wheres-the-fun/round5-mellanie.md @@ -0,0 +1,494 @@ +# Round 5: Mellanie — Copy Perspective on 15 Decisions +**Workshop:** Where's the Fun? | **Round:** 5 (Deep Dive) | **Agent:** Mellanie + +--- + +## Framing: What These 15 Decisions Do to My Domain + +These 15 decisions collectively change copy work more than any other single round. Three changes are architectural: + +1. **Career-to-culture voice inversion (D6)** — Voice cards authored at culture level, not job level. The architecture of what I've been building is inverted. +2. **Named NPCs to generated NPCs (D7)** — Every specific authored line referencing Kael, Naia, Maret becomes either a template slot or dead stock. +3. **Tycoon as v0.2 bookmark (D4)** — All smuggler/detective voice work is now deferred. The tycoon is a blank. + +The good news: the *delivery structure* is clearer than it's ever been. The first authored moment is defined (alarm clock → apartment → insert activation). The setting delivery mechanism is confirmed (insert copy as parallel layer to visual). The voice model has a direction. What's missing is enough specificity to start writing production content. This round flags every rough edge I can see and asks for the decisions needed to unlock the copy pipeline. + +--- + +## Decision 1: Proof-of-Life = Generator + Graphics + +The PoL is the generator running at scale plus legible characters. No hand-built slice. No authored scenes. + +**For copy, this means:** My deliverables don't block or define the PoL. No dependency runs from my domain to the PoL milestone. This is clarifying. + +**The risk:** Copy gets treated as garnish you add after the technical foundation ships. It isn't. The insert activation (D13) and the first morning monologue (D14) are part of the first-run experience even if the PoL doesn't include them. If the PoL succeeds and we move immediately to "build the game on top," the copy pipeline needs to be ready to engage fast. I should not be the bottleneck after the generator proves out. + +**Recommendation:** Use the PoL sprint to spec the voice model, not to write content. When the generator ships, the copy pipeline is already designed and ready to produce. The sequence: design voice architecture → spec insert copy categories → wait for generator proof → write lines. + +--- + +## Decision 2: Skills + Bookmark Only for Character Creation + +Character creation has two inputs: skill budget and bookmark (career). No culture, no religion, no family in v0.2. + +**For copy:** This is workable, but it creates a direct tension with D6 (see below). Skills can carry some voice weight — a high-social tycoon notices power dynamics in every room; a high-mechanical tycoon notices the HVAC system — but skills are a modifier, not a register. The base register needs to come from somewhere, and in v0.2 the culture layer doesn't exist yet. + +**The workable part:** Two skill-aware layers on the tycoon voice card, not a separate card per skill profile. "This tycoon notices social dynamics" vs "this tycoon notices systems" are monologue variants within one culture-baseline card, not separate voices. + +--- + +## Decision 3: Religion Is Not a Game System + +One fewer voice modifier. Clean. No ritual language, no faith-adjacent idioms unless worldbuilding develops them organically. + +This removes a potential source of cultural specificity that I might have drawn on for voice distinctiveness. I'll park it. The Settled Reach's economic texture (Commission hierarchy, span gate access tiers, district class stratification) provides enough cultural specificity for voice differentiation without religion. + +--- + +## Decision 4: Tycoon Is the v0.2 Bookmark — Zero Investigation + +**The biggest copywriting shift of Round 4.** + +My smuggler voice card is on hold. My detective voice card is on hold. The tycoon is a blank. And I mean that literally — I have no character brief for what a tycoon in the Settled Reach sounds like from the inside. + +The smuggler I understand. Pragmatic, quietly defensive, warm with people they trust, uncomfortable with direct confrontation, good at not asking the question they know the answer to. I understand the smuggler because the moral arc forced me to understand who that person *is*. + +For the tycoon: I don't know yet. And this matters urgently. + +**Tycoon-specific voice questions I'm working through:** + +- **Visibility.** Does the tycoon want to be seen? Smugglers stay invisible. Most tycoon archetypes are visibility-maximizers — reputation is currency. If that's true in the Settled Reach, the interior monologue has a fundamentally different relationship to public vs. private self than the smuggler's. +- **What they notice.** A social-manipulation-heavy tycoon reads power dynamics instantly — they walk into a room and the monologue names who has leverage over whom before the player consciously registers it. A mechanically-inclined tycoon notices inefficiencies, bottlenecks, systems. Either way, the tycoon reads the world through an economic lens. +- **What they fear.** In a world with Commission oversight and span gate access hierarchies, the tycoon might fear visibility *of the wrong kind* — not the Commission's moral judgment, but their attention. They've built something worth protecting, which creates a specific kind of paranoia. Different from the smuggler's fear (getting caught). Closer to: what I've built can be taken. +- **Their relationship to the insert.** This is a key voice question. The insert as a business tool means the tycoon's internal monologue might blur with the insert's informational overlay — thinking in market data, in contact hierarchies, in deal structures. The insert isn't a tool they pick up; it's how they see. That blurring is a voice characteristic. + +I can speculate. But I need a design anchor before I write production content. **See Q1.** + +--- + +## Decision 5: Skills Affect Outcome (Mostly C) + +Everyone sees the same verbs. Skills determine how well you do. Some advanced verbs may still be gated — spec needed for which ones. + +**For copy:** The trigger catalog simplifies at one end and complicates at another. + +Simplifies: no lines for "you can't do this because your skill is too low" (verb-gated content). The verb is always visible. I don't need to write inaccessibility. + +Complicates: I need **outcome-differentiated lines**. The character attempting an action and executing it well is a different interior experience from the character attempting the same action and executing it poorly. A high-social-manipulation tycoon who lands a negotiation cleanly has a specific thought about it. A low-social-manipulation tycoon who botches the same negotiation might not even know what went wrong. + +**New trigger I'm proposing:** `verb_outcome` with a quality parameter — `verb_outcome_high`, `verb_outcome_low`, `verb_outcome_catastrophic`. The interior line at outcome differentiates by skill result, not by verb availability. The player chose *Talk*; the character experienced the consequences of how well they talked. + +This makes skill results visible through interiority without being a tutorial. The player learns their character's skill profile by how the character *feels* about their own performance — not by a success/failure popup. + +Needs coordination with Gestalt on which verbs will commonly be attempted at low skill (those are the verb_outcome_low lines I prioritize writing). + +--- + +## Decision 6: Voice Is Culture-Driven, Job Modifies + +**This answered my Round 4 Q1 — but only partially, and with a new problem attached.** + +Jeroen confirmed: culture is primary, job adds a layer. The character IS their background. Job is what they do with it. A Krenn tycoon sounds like a Krenn person running businesses, not a generic tycoon. This is the right architecture — it produces a richer voice space and explains why two people with the same job might read completely differently. + +**The new problem:** Culture is deferred from v0.2 character creation (D2). Skills + bookmark only. So in v0.2, the culture layer doesn't exist. + +This creates an authoring gap: Decision 6 says voice is culture-driven, but v0.2 won't have culture data to drive it with. + +**Three options:** + +**Option A — Default culture baseline:** Choose one culture as the implied default for all v0.2 characters. Write the tycoon voice card for that culture + tycoon modifier. When culture selection ships, other culture voice cards join it, and the tycoon modifier applies to all. Cost: one culture's assumptions baked in as defaults. Players who conceptualize their character differently will feel the mismatch. + +**Option B — Skill-based register bridge:** Use skills as a temporary voice differentiator. High Social → warmer, more relational register. High Technical → more analytical, systems-thinking. Skills fill the culture role until culture ships. Cost: this is backwards from D6 (culture primary, not skills) and creates content that needs rearchitecting when culture arrives. + +**Option C — Deliberately parametric tycoon voice:** Write the v0.2 voice card with culture slots explicitly open — templated register parameters that culture will fill later. Write the tycoon modifier as a modifier applied to a culture template. Less sharp content now; correctly architected later. + +**My recommendation is C** — slightly blander v0.2 content, but correctly structured for what culture adds later. Option A bakes in a default that becomes the reference all future culture variants are compared against; get it right or it's permanent debt. + +**Question for Jeroen: See Q1.** + +--- + +## Decision 7: All NPCs Generated, No Named Characters + +**Paula's domain takes the biggest hit from this decision. Mine takes the second biggest.** + +Kael doesn't exist. The generator produces NPCs that fit positions based on location characteristics. All authored content about named characters — the Kael voice moments, the Naia observations, the Maret behavioral tell — is either a template to be parameterized or dead stock. + +This is the right call. The life-sim framing requires it. Rimworld colonists become specific and beloved through player experience, not through authored backstory. + +**What this destroys in the current copy canon:** + +Every line that references a proper noun is dead as production content. "Kael's already at the dock. Good." "She used to just sign. Now she reads every line." These reference authored history that won't exist in the generator model. They're now register demonstrations — examples of the emotional texture and approach — not line templates. + +This is a necessary pipeline clarification: audit all existing voice examples and tag them `[register: demonstration]` or `[template: valid]`. Most existing named-NPC lines are the former. They're still useful — "this is how a Krenn smuggler sounds when they have years of history with someone" — but the name is a slot, and the specific behavioral history is a delta the knowledge graph would need to surface. + +**The deeper problem: intimacy without history** + +The smuggler's Phase Zero warmth was built on embedded backstory — years of shared shifts implied in every line. With generated NPCs, there is no authored backstory. The character meets their FRIEND-equivalent the way you meet anyone: fresh. + +This is a different emotional register. The warmth can't be "already familiar." It has to be "becoming familiar." Three stages: + +1. First observation of this NPC: professional appraisal. *"Three shifts with [Name] now. They work fast and they don't need managing. That's most of what I need."* +2. After repeated contact: beginning to notice patterns. *"[Name] always counts the manifest twice. Once fast, once careful. Something in how they were trained."* +3. Warmth established: genuine recognition. *"[Name]'s in early. Good sign or a bad one — hard to tell yet."* + +The trigger system needs to distinguish these relationship states. The monologue has to respond differently at each. This is more complex than the authored-NPC system, but more *honest* — the player builds the warmth themselves rather than discovering warmth the author pre-loaded. + +**The behavioral delta problem:** + +The intimacy of authored-NPC monologue came partly from knowing the before-state. "She used to just sign" requires knowing prior behavior. With generated NPCs, the simulation tracks behavioral changes — but the monologue needs the *delta*, not just the current state, to write the equivalent intimacy: + +> *"More careful at the manifest today. Wasn't like that last week."* + +This requires the knowledge graph to surface behavioral change from baseline, not just current state. It's a systems ask — for Gestalt and Tyre — but it's what makes generated-NPC monologue carry the same intimacy as authored-NPC monologue. + +**Question for Jeroen: See Q2.** + +--- + +## Decision 8: Generative AI for NPC Content Templating + +Culture vectors, tone, accents as templating dimensions. Limited vocabulary acceptable at first. In-game ollama for live dialogue deferred but on the table. + +**The opportunity and the risk are exactly as large as each other.** + +The opportunity: the copy pool for a generated NPC world is enormous. Culture × tone × accent × situational variant × career modifier = a volume of content that's not hand-authorable at any reasonable team size. AI-assisted templating is the right production model. + +The risk: AI-generated content at scale with no quality bar sounds like AI-generated content. The uncanny valley for prose isn't visual — it's tonal. A world full of NPCs who speak in slightly-off registers, who lack idiosyncrasy, who sound like they're completing a prompt rather than being a person — that world feels hollow even if the simulation underneath is rich. + +**The new competency this requires:** + +1. **Prompt specification writing.** The voice card becomes a prompt — and prompt writing is different from line writing. "Fragment sentences are not a mistake, they're the voice" is a voice card note. In a prompt, it becomes: "Always use short declarative fragments for internal monologue. Maximum 8 words unless it's an operational calculation. Never write complete subject-verb-object sentences unless the character is making a formal assessment." Same intent, different form. + +2. **Quality criteria for AI output.** What does it mean for an AI-generated line to pass the voice card? The checklist items that make a human author's line good also need to be evaluable on generated output. "Does this sound like they're thinking, or like they're explaining?" is a judgment call that requires a trained ear applied to batches of 50 candidate lines. + +3. **Anti-pattern detection at scale.** The voice card's anti-pattern list (no self-analysis, no abstract emotional states, no complete sentences in fragment-register voices) needs to be checkable across batch output. Some anti-patterns are trivially detectable ("I feel relieved" → flag immediately). Others are subtle (a line that's the right length but uses passive voice inappropriately for this character's register). + +**The quality floor concern:** + +"Limited vocabulary acceptable at first" applies to NPC dialogue. I want to make sure this doesn't extend to player character monologue by accident. + +NPC dialogue with limited vocabulary: acceptable. The Sims proves this. "Hello." "Nice to meet you." — repetitive, but it works because emotional content comes from behavior (the animation, the relationship state), not from words. + +Player character monologue with limited vocabulary: not acceptable. The character's interior voice is the most intimate channel in the game. If the player hears the same 12 monologue lines cycling, the character stops feeling like a person and starts feeling like a recording. The intimacy collapses. + +The distinction needs to be explicit in the content plan: NPC dialogue gets AI-assisted limited-vocabulary approach first. Player character monologue gets hand-authored or heavily curated lines for core triggers, AI assist for edge cases and high-volume situations. + +**Question for Jeroen: See Q3.** + +--- + +## Decision 9: Possible In-Game Ollama for Live NPC Dialogue + +Deferred but the door is open. + +If live dialogue via ollama is on the roadmap, the authored content layer isn't the ceiling — it's the scaffolding the LLM learns from. Authored lines establish the pattern; the LLM extends it dynamically. + +**What this means for my work now:** Don't write for a closed vocabulary. Write exemplar lines that could teach a model what this voice sounds like. The authored content becomes training signal, not final output. That's a reason to write with even more intentionality about what makes each voice distinctive — the exemplars are teaching something. + +Every voice card line I produce for the culture/tycoon register should be written as if it's demonstrating the quality bar, not just filling the content pool. This changes the mentality slightly but not the craft. + +--- + +## Decision 10: Quietly Responsive World — Gradient of Caring + +The world doesn't care globally but notices locally. Primary social contacts (colleagues, neighbors) develop responsiveness over time. Gradient of caring based on social proximity. + +**For copy, this gradient has trigger implications:** + +The relationship-proximity level of the NPC being observed or referenced should be a trigger parameter. `relationship_proximity: stranger | acquaintance | colleague | primary_contact` affects which line pool fires. A character watching a stranger walk by is in ambient observation mode. A character watching their primary contact walk in is in warmth-recognition mode. Same verb (*observe*), different interior. + +Watching a stranger: +> *"Commission cap. They always stand with their feet apart like that."* + +Watching an acquaintance: +> *"Darros again. He's early — something's moved."* + +Watching a primary contact: +> *"[Name]. They're in early. Good sign or a bad one — hard to tell yet."* + +Three observers, three versions of "I see someone I know at various distances." The monologue system needs to distinguish these proximity states to deliver the right line. This is not a new insight — Paula and I have discussed the trigger architecture for this — but D10 confirms it's part of the designed experience, not optional flavoring. + +**D10 also resolves my Phase Zero concern:** I'd been uncertain whether the world was so uncaring in early build that Phase Zero warmth would feel incongruous. Jeroen's answer clarifies: the world is never Kenshi-indifferent, even in Phase 1. Colleagues and neighbors develop responsiveness. That's exactly the soil Phase Zero warmth grows from. I don't need to wait for Phase 2 authored arcs — ordinary life, ordinary contact, ordinary accumulation of recognition is the content. + +--- + +## Decision 11: Full Character Customization + +Hair, clothing, colors. The creation screen is part of identity investment. Readability at tile scale solved through outline/highlight. + +**For copy:** Monologue lines must be appearance-agnostic. No lines that assume hair color, height, specific physical tells. The character's physical self-image is the player's choice; I can write interiority without referencing it. + +One positive implication: the creation screen's identity investment means players are emotionally invested in their character before the first line fires. The first morning monologue is heard by a player who has already made choices about who this person is visually. The line has a face to land on. That's a gift — it means the monologue doesn't need to do as much heavy lifting to establish "this is a specific person" at Day 1. The player has already done some of that work. + +--- + +## Decision 12 + 13: Insert Copy as Setting Delivery and First Moments + +**The decisions:** Setting delivered through both visual and insert. The first two moments are waking up in your auto-generated apartment and insert activation. Both are intimate and personal before the player leaves the room. + +**The insert is a new copy register I haven't worked in yet.** + +The insert is not the character's interiority. It's a tool. It has data. It has system messages. It might have personality (a tycoon insert configured to speak in financial analyst language; a law enforcement insert with Commission bureaucratic diction). But it is not the character's feelings — it's the character's interface. + +Insert copy register: **data labels with embedded worldview.** + +The insert tells the player what kind of world they're in by deciding what to label, how to label it, and what to flag as noteworthy. The same location through two different inserts: + +Tycoon: +``` +THE LAST SHIFT (bar/transit-adjacent) +Estimated asset value: 14,200 cr +Current owner: Pell Darros +Location premium index: 2.1x (underpriced for adjacency) +Foot traffic: HIGH (span gate secondary) +Note: flagged for acquisition opportunity +``` + +Law enforcement: +``` +THE LAST SHIFT +Classification: Licensed hospitality, unrestricted +Compliance status: CURRENT (last inspection 14 days ago) +Known associations: regular foot traffic, mixed +Incidents logged: 2 (minor, unresolved) +``` + +Same world. Two completely different emotional registers. The tycoon sees opportunity and underpricing. The law enforcement sees incidents and associations. This is setting delivery that is inherently diegetic, inherently career-specific, and requires no exposition. It's good design. But it requires two separate insert copy registers for two separate career paths, and more for each career added later. The copy work for insert is significant and hasn't been scoped. + +**The apartment first moment:** + +Waking up before the insert activates. The first monologue line of the game fires here. + +What this line must do: +1. Establish that you're in YOUR space (not neutral, not hostile) +2. Establish the character's relationship to mornings (operational? reluctant? calculated?) +3. Establish economic register without stating it (a wealthy apartment reads differently than a sparse one) + +The apartment's auto-generated economic status needs to signal to the monologue system so the first-morning line fires from the right economic register. New trigger: `apartment_wakeup + economic_tier`. + +For a lower-tier apartment (span gate noise, recycled air): +> *"The span gate recycled three times before the alarm. Could hear it through the wall."* + +For a higher-tier apartment (quiet, self-scheduled): +> *"Seven-thirty. I didn't set an alarm."* + +Neither announces its class position. Both communicate it through specific detail. + +**The insert activation moment:** + +The insert powers on. Neural implant boot sequence. Intimate and slightly technical — like your phone booting up, except inside your skull. + +The boot sequence is where the insert's personality and diction establish themselves. If the tycoon has configured their insert voice, the activation message sets that register. This is the player's first encounter with a copy register they'll see every session. It needs to feel: + +- Technological but not cold — this is part of your body +- Personalized but not cloying — the insert knows you but isn't performing warmth +- World-specific — "SOVA TRANSIT NEURAL MESH" or similar names this world in the first second + +What it should NOT say: "Welcome back, user." That's a UI placebo. The insert in this world is intimate enough to have a relationship with. It knows your priorities already. It surfaced the right message first. + +**I need to develop an insert copy style guide** before the first activation lines are written. The style guide answers: what is the insert's relationship to the character? How does it present data? When does it annotate vs. just report? What's the tycoon configuration vs. the default vs. the law enforcement configuration? See Q4. + +--- + +## Decision 14: Groundhog Day Alarm Clock Homage + +*Click pa-pa pa-pa, cut short. First game day only. "New day, new start, new chances" — with a wink.* + +The audio design is Quentin and Cass's domain. My domain: what text accompanies that moment, and what does the character's interior experience of waking up sound like? + +**The wink is a tonal anchor.** It means the game knows what it is. Not grimly realistic, not earnestly serious. There's lightness. The character who wakes up to this alarm can be wry, can have a beat of self-awareness that the later, heavier content won't have. + +The first morning monologue line is the hardest line in the game to write. Everybody hears it. It sets the register. It has to be specific, slightly wry, real — the character awake and present in their own life, not performing for the player. + +**Day 1 specifically:** The character knows this is Day 1 of a new chapter. Different line from Day 47. The Groundhog Day homage is audio; the copy companion to that moment should carry possibility, not weariness. + +**Speculative Day 1 tycoon line (voice not yet confirmed):** +> *"New district. First meeting at nine. The insert's already mapped my route — two transit passes, contingency if the span gate backup queue runs long. I put in the skill points for negotiation. Today we'll see if I put them in the right place."* + +Right shape. Wrong certainty about voice. See Q1 for the design conversation needed. + +**After Day 1:** The `morning_wakeup` line pool needs 15-20 variants to cover the first month of play without repetition. Variants can respond to: economic tier, previous session's outcome (deal won/lost), relationship state (primary contact is in good standing / troubled), and simple rotation. This is a high-priority content deliverable — it fires every single session. + +--- + +## Decision 15: Player Choices ARE the Content (Rimworld Model) + +One authored starting beat (alarm clock + first appointment), then agency and options. Job is rails to take off from, not a script to follow. + +**The trigger catalog must be wide.** The player can go anywhere and do anything from minute five onward. The monologue needs something to say about all of it. + +Wide trigger coverage, not deep authored paths. Instead of a rich authored arc from beat 1 to beat 30, I need shallow-but-comprehensive coverage: `enter_location` lines for every zone type, `observe_npc` lines for every archetype type, `job_event` lines for every common tycoon event, `consequence_visible` lines general enough to fire on any consequence type. + +The risk of this model: if the player does something the trigger catalog doesn't cover, they hear nothing. Silence where the character should have a thought. This is worse than a generic line — a generic line at least signals presence. Silence signals "you've left the authored space." + +The solution is the same as Rimworld's: write to archetypes, not specifics. Broad trigger categories with enough line variants that the match feels tight even if it isn't exact. + +**Tycoon-specific trigger categories for a minimum viable catalog:** + +| Trigger | Notes | +|---------|-------| +| `morning_wakeup + economic_tier` | 15-20 variants, daily fire | +| `enter_location: terminal/dock` | Operational, transactional | +| `enter_location: bar/hospitality` | Social, opportunity-reading | +| `enter_location: market_district` | Economic assessment | +| `enter_location: commission_offices` | Careful, aware | +| `enter_location: residential` | Contextual, status-reading | +| `observe_npc: colleague_warm` | Recognition at proximity | +| `observe_npc: business_contact` | Opportunity assessment | +| `observe_npc: hostile_competition` | Threat appraisal | +| `observe_asset: own/performing` | Satisfaction, ownership | +| `observe_asset: own/underperforming` | Concern, calculation | +| `observe_asset: unowned/flagged` | Opportunity | +| `job_event: contract_offered` | Appraisal | +| `job_event: contract_accepted` | Commitment | +| `job_event: contract_completed_success` | Satisfaction | +| `job_event: contract_completed_partial` | Recalibration | +| `job_event: contract_failed` | Assessment, not catastrophizing | +| `verb_outcome_high` | Skill landed | +| `verb_outcome_low` | Skill missed | +| `relationship_shift: acquaintance→colleague` | Recognition | +| `relationship_shift: trust_broken` | Recalibration | +| `consequence_visible` | Downstream effect noticed | +| `economic_event: market_shift` | Reading the change | + +Twelve lines minimum per trigger; twenty-four is better. This table is the minimum viable tycoon monologue spec. Every trigger without coverage is a silence the player will eventually notice. + +**The Rimworld model also resolves my Phase Zero concern from Round 4.** I'd framed Phase Zero as a pre-content warmth-building period before authored arcs inject. The Rimworld model says Phase Zero *is* the content — the ordinary life IS what happens, not a prologue to what happens. That means ordinary life monologue (the wide trigger catalog above) is the primary content deliverable. Dramatic arc content (Phase 1 crack, Phase 2 consequence) layers on top of a well-developed ordinary-life base. + +Write ordinary life first. Write drama after. This is the right sequencing. + +--- + +## Cross-Agent Connections + +**With Miri (D12 + D13):** The insert copy can't know what to communicate if the worldbuilding layer hasn't designed what's true about this district and economy. A tycoon's morning overlay mentioning "Docking Fee Subsidy Expiring in 14 Days" requires Miri to have designed that Sova Transit has docking fee subsidies. The zone identity spec Miri is recommending is upstream of my insert copy. I cannot write the morning overlay until Miri has written the district's economic texture. + +Three-way handoff needed between Miri (worldbuilding), Araminta (insert visual grammar), and me (insert copy) before first-frame copy is finalized. These aren't parallel tracks — they're sequential: worldbuilding → visual grammar → copy fits the grammar. + +**With Paula (D7 + D10):** Paula's Phase Zero warmth content and mine converge on the same problem from different angles. Paula is writing the arc structure; I'm writing the lines that carry it. Both of us are now building for generated NPCs rather than named ones. + +Our joint deliverable: a template system for warmth-building monologue that fires as the relationship metric climbs, using the player's NPC's name and role rather than Kael's. The Phase Zero content architecture Paula and I sketched still applies — it just operates on generated NPC slots. We need a working session to convert the named-NPC content spec into a template spec before either of us writes production lines. + +Paula's point from Round 4 about the visual-content handoff is right: Phase Zero warmth lines need to know what the generated FRIEND-NPC looks like and how they move. A line like "[Name] makes that face when customs is watching" works if the player can read the NPC's face. If NPC legibility isn't solved, this line registers as noise. Araminta's work comes before Paula's and mine, in this specific dependency chain. + +**With Gestalt (D5):** The `verb_outcome` trigger I'm proposing needs coordination with the VerbPriorityProfile spec. I need to know which verbs are commonly attempted at low skill — those are the verb_outcome_low lines I prioritize. If the tycoon's skill profile emphasizes social manipulation, I write more social-failure lines. If mechanical repair is low, one or two failure lines for those situations, not a full pool. + +**With Araminta (D11 + D12):** The insert visual grammar is the container for my insert copy. I can write insert text that works in any visual context, but it'll be better if I know how the insert presents information — AR overlay text? A panel? Distinct informational zones (contacts / calendar / market / alerts)? Copy structure should fit visual structure, not be retrofitted after the fact. + +--- + +## Three Copy Near-Misses to Avoid + +### Near-miss 1: Writing job-level voice cards when culture is the primary driver + +D6 (culture primary) + D2 (culture deferred) = risk of writing "tycoon voice card" as if tycoon is the base, then having to retrofit culture architecture when it ships. Mitigation: write the tycoon voice card as a modifier template explicitly, with culture slots marked as open. Less sharp now; correctly architected later. + +### Near-miss 2: Writing named-NPC lines as production content + +The existing voice exemplars are register demonstrations, not production templates. Every line referencing Kael, Naia, Maret by name is dead as production content. It's still useful as a register target — this is how the Krenn smuggler sounds when talking about someone they've worked with for years — but the name is a slot, and the specific behavioral reference requires a delta the knowledge graph has to surface. + +Audit all existing voice examples and tag them `[register: demonstration]` or `[template: valid]`. Most existing named-NPC lines are the former. + +### Near-miss 3: Applying "limited vocabulary acceptable" to player character monologue + +Jeroen said this in context of NPC dialogue. The player character's internal voice is the primary intimacy channel. If that voice cycles through the same 12 observations, the character stops feeling like a person. NPC dialogue: limited vocabulary, AI-assisted templating, acceptable repetition. Player character monologue: hand-authored or heavily curated, broader coverage, repetition is a defect. + +This distinction needs to be explicit in the content planning. It's easy to say "AI-assisted pipeline" and have it apply uniformly to everything. It shouldn't. + +--- + +## Questions for Jeroen + +### Q1: What IS a tycoon in the Settled Reach — and what drives voice in v0.2 when culture is deferred? + +Two questions entangled. + +**The character question:** I know what a smuggler is, emotionally and psychologically, because the moral arc forced me to understand who that person is. The tycoon is a blank. Before I can write a tycoon voice card line with confidence, I need to understand: + +- Does the tycoon want to be seen, or do they operate through selective visibility? +- What are they afraid of in this specific world — Commission scrutiny? Competitor intelligence? Losing narrative control of their own reputation? +- What is their relationship to the insert — trusted partner, necessary tool, something slightly uncomfortable about how much it knows about them? +- What's the emotional register — calculating but warm? Ambitious but anxious? Confident but always aware they're one bad deal from falling? + +The Settled Reach tycoon should be specific to this world and its social texture. What does economic aspiration feel like when you're operating below the span gate access tier? When the Commission is a constant ambient fact? A paragraph from Jeroen on "who is the tycoon" would unlock weeks of content work. + +**The voice architecture question:** D6 says culture-driven voice. D2 says culture is deferred. For v0.2, do I write the tycoon voice card as: (A) one culture's baseline explicitly named, (B) a deliberately parametric template with culture slots left open, or (C) skills-as-temporary-voice-differentiator? Option B is my recommendation. But "what culture baseline fits the tycoon" requires Jeroen's call on the Settled Reach's economic sociology. + +--- + +### Q2: Are behavioral deltas surfaced to the monologue system, or only current behavioral state? + +Authored-NPC monologue carried intimacy partly from knowing the before-state. "She used to just sign" is a contrast — it requires knowing prior behavior. With generated NPCs, the simulation tracks behavioral changes, but the monologue needs the *delta* from baseline to write equivalent intimacy: + +> *"More careful at the manifest today. Wasn't like that last week."* + +This requires the knowledge graph to surface behavioral change from baseline, not just current state. Is this a system capability the simulation team can support? If yes, the generated-NPC monologue can carry the same intimacy as authored-NPC monologue. If no, the lines have to work only from current state — which is a different (weaker) kind of interiority. + +This is a systems question as much as a copy question. I want to flag it now so it's on Gestalt's and Tyre's radar when they design the knowledge graph. + +--- + +### Q3: Is player character monologue explicitly excluded from "limited vocabulary acceptable at first"? + +Jeroen said this in context of NPC content. I want to confirm explicitly that it doesn't extend to the player character's interior voice. + +NPC dialogue with 15-20 variants per state: acceptable. The simulation and behavioral signal carry the emotional content; words are secondary. + +Player character monologue with 12-15 cycling lines: not acceptable. The monologue is the character. Repetition destroys intimacy. I'd rather ship tycoon content with correct but limited trigger coverage (every trigger covered, 12 lines each) than wide coverage with shallow pools (every trigger covered, 5 lines each, player notices cycling in 30 minutes). + +Confirming this distinction shapes how I allocate authoring time. If the quality bar for player monologue is higher than for NPC dialogue, I write fewer triggers with deeper pools rather than more triggers with shallower pools. + +--- + +### Q4: What is the insert's relationship to the character — and does it have a "voice"? + +The insert is a neural implant the character has had for years. It knows their preferences. In the tycoon case, it surfaces market data, contact messages, deal notifications. It might surface things in the order the character cares about, not in chronological order. + +Does the insert have a configured "personality" — a register the character chose when they set it up? Or is it a neutral data terminal? The difference: + +A configured-personality insert is a character in its own right — slightly uncanny, intimately knowing, its first boot message establishing a relationship. A neutral-data insert is a UI element that reads more like a heads-up display. + +My instinct is configured personality — it makes the technology feel intimate and world-specific rather than generic sci-fi HUD. But this is a design decision, not a copy decision. The answer shapes the entire insert copy style guide. + +--- + +## What I'm Doing and What I Need + +**What I'm doing:** +- Speculating on the tycoon voice register to have a draft ready when the design conversation happens +- Speccing insert copy categories for v0.2 scope (morning activation, calendar overlay, deal notification, relationship ping) so I can write to a defined target once voice is confirmed +- Drafting the template conversion rules for named-NPC content (how Kael-specific lines become FRIEND-role templates) so Paula and I can align on Phase Zero warmth architecture +- Developing the minimum viable trigger catalog (see D15 section) so it's ready to review + +**What I need before writing production content:** +1. Answer to Q1 (tycoon character brief + culture baseline for v0.2) +2. Answer to Q4 (insert voice design) +3. Three-way handoff with Miri and Araminta on first-frame worldbuilding → visual grammar → insert copy +4. List of v0.2 insert contexts in scope (what states does the tycoon's insert enter?) +5. Answer to Q2 from Gestalt/Tyre (behavioral delta surfacing) + +**What I'm not doing until the above are answered:** +- Writing tycoon monologue lines for production +- Writing insert activation copy for production +- Building the NPC template pool + +--- + +## My Single Most Important Recommendation + +**Write the tycoon character brief before any tycoon content.** + +Not the voice card. Not the lines. The brief. One to two paragraphs answering: who is this person? What do they want, notice, fear, reach for? What makes the Settled Reach tycoon specific rather than generic "ambitious businessperson"? + +The smuggler arc was emotionally powerful because it was built on a specific human situation — someone in the gray zone, warming to a person who is also in the gray zone, not yet knowing what that warmth will cost. The tycoon needs an equivalent specific situation. My candidate: someone trying to build something real in a world where the span gate separates the successful from everyone else, working below the tier they're aiming for, watching the Commission's presence as both threat and potential tool, knowing that one bad deal or one wrong contact collapses everything they've constructed. That's a character. A tycoon voice card written from that character is interesting. A tycoon voice card written from "ambitious business person" is a genre placeholder. + +The brief doesn't need to be long. It needs to come from Jeroen, because the emotional specificity of the Settled Reach tycoon is a design decision, not a copy decision. + +Once I have the brief, I can have a first draft voice card within a day. That voice card unlocks the morning lines, the insert activation copy, the verb-outcome lines, the warmth-building templates — the entire content pipeline moves. + +Without the brief, I'm building on a guess. And the v0.1 lesson is: don't build on guesses. + +--- + +*Mellanie — Round 5 complete.* diff --git a/docs/workshops/wheres-the-fun/round5-miri.md b/docs/workshops/wheres-the-fun/round5-miri.md new file mode 100644 index 000000000..bf6fefc65 --- /dev/null +++ b/docs/workshops/wheres-the-fun/round5-miri.md @@ -0,0 +1,483 @@ +# Round 5 — Miri: Reactions to 15 Decisions +## Where's the Fun? Workshop | 2026-03-05 + +**Domain:** Worldbuilding & Setting Design +**Reading:** Round 4 interview transcript, Round 4 decisions supplement, confirmed against existing D-records + +--- + +## Preliminary: Lore Check + +The interview uses "Krenn tycoon" as a voice example. Confirming this against established records: "Krenn" is our confirmed star system (D-036, D-050), not a species or distinct cultural group. The Krenn System is the mid-Reach G3V system where Station Sova orbits Velen. NPC naming conventions are Krenn System conventions — compact, consonant-heavy, first-name-primary (Kael, Voss, Naia, Pael, etc.). So "Krenn tycoon" = a tycoon from the Krenn System, in the same way you might say "a Londoner." This is consistent with our existing lore. + +What this reveals: "Krenn" as a cultural identity is regional, not ethnic. It's a place that produces a type of person. That's a worldbuilding model I can work with — and it's important, because it means culture in this game is tied to *location* and *history* as much as anything else. A person who grew up on Station Sova speaks and thinks differently from someone who grew up in a core-system transit hub. This is exactly right for the life-sim model. + +However: the Settled Reach has many star systems. If the generator is eventually producing multiple locations, those locations will produce different cultures. For v0.2, the scope question is whether the generator is producing multiple Settled Reach locations or just Sova Transit / the Krenn System at scale. That determines how many cultures I need to design for the first content pass. + +--- + +## Reactions to the 15 Decisions + +### Decision 1: Proof-of-life = generator + graphics, not hand-built slice + +**Setting note — STRONG ENDORSEMENT, with urgency flag.** + +This is the correct call, and it validates my Round 4 argument that the zone identity spec is not polish — it's foundation. But it also raises the stakes significantly. + +v0.1's "game level with dots" problem happened with a HAND-BUILT location. Jeroen, Paula, Tyre, and others spent sprints crafting Sova Transit manually, and it still read as generic. A generator that produces locations without zone identity rules will produce the same failure at scale — generic space procedurally stamped out, infinite "game levels with dots." + +The generator must know what the Settled Reach looks like before it runs. This means: +- Zone type taxonomy → zone identity rules is the dependency chain +- Zone identity rules must exist before the generator proof-of-life is demonstrated + +If the generator runs for v0.2 and produces legible Settled Reach space, we've proven two things at once: the technical foundation AND the worldbuilding expression layer. If the generator runs and produces generic space, we've proven nothing that v0.1 didn't already falsify. + +**My commitment:** The zone identity spec is now my v0.2 Priority 1 deliverable, not "one sprint after other setup." It needs to exist before the generator is tested. + +--- + +### Decision 2: Skills + bookmark only for character creation (culture deferred) + +**Setting note — TENSION FLAG. This decision is in conflict with Decision 6.** + +Skills + bookmark is a reasonable scope constraint. But "culture deferred" creates a problem I need to name clearly. + +Decision 6 says culture is the PRIMARY driver of voice. Decision 7 says the generator produces NPCs using cultures as a key variable. Decision 8 says the generative AI pipeline uses culture vectors, tones, and accents as templating dimensions. + +All three downstream decisions depend on culture as a designed game object. Yet culture is deferred from character creation. + +This could mean one of three things: +1. **Culture is implicit in bookmark.** The tycoon bookmark in the Krenn System automatically implies Krenn culture. Character creation doesn't need a culture field because location + history determine it. This is elegant and worldbuilding-consistent (culture IS place + history in this model). +2. **Culture is selected but not called "culture."** Maybe the character creation screen has a "background" or "upbringing" field that produces culture without naming it as such. +3. **Culture is undefined for the player, defined only for NPCs.** The player has no cultural identity; only the generated world does. This would mean the player's voice card doesn't have a culture base — which contradicts "a Krenn tycoon sounds like a Krenn person who runs businesses." + +I need to know which of these is intended before I can write culture profiles. My preference is option 1 — culture is implicit in starting location — because it keeps the design simple and is worldbuilding-consistent. But I need Jeroen to confirm. + +**Flag:** "Culture deferred" cannot mean "culture undefined." The voice system, NPC generator, and AI pipeline all need culture as a working concept immediately. + +--- + +### Decision 3: Religion is NOT a game system + +**Setting note — noted and clean.** + +Religion exists in the Settled Reach as lore (people believe things, have traditions, have cultural practices shaped by history). It's just not a gameplay mechanic. This is the right call for this game — the moral texture comes from economic and social choices, not theological alignment. + +No concerns. I'll continue treating religious practices as NPC background flavor (consistent with the "quotidian-with-undertow" atmosphere of D-036) without designing any mechanical faith systems. + +--- + +### Decision 4: Tycoon is the v0.2 bookmark, zero investigation + +**Setting note — ENTHUSIASTIC ENDORSEMENT, with a worldbuilding opportunity note.** + +From a setting design perspective, this is actually richer than the detective bookmark. Here's why: + +The detective sees Sova Transit as a surveillance problem — who is where, what are they hiding. The tycoon sees Sova Transit as an economic map — what is worth owning, who controls what, where the value flows. The tycoon's perspective reveals the district's social geography in a different way: the logistics hub becomes a revenue-generating operation with Commission licensing fees. The bar becomes a property with captive clientele and potential for social leverage. The corridor zones are transit arteries with adjacent storage that commands rent. + +The tycoon also has a natural relationship to the class anxiety baked into Sova Transit. Span gate access is expensive — maybe the tycoon's goal is to build enough capital to buy access and leave for higher-system work. Or maybe they realize that owning businesses on a transit hub means span gate traffic is their customer base, not their aspiration. + +**What Sova Transit needs for the tycoon bookmark specifically:** +- A property market (what can be owned, rented, bought into) +- An economic layer (what things cost, what revenue streams look like) +- A Commission relationship (licensing, inspections, fees — the friction of legitimate business) +- A social geography map from the tycoon's perspective (which NPCs are potential employees, customers, rivals, contacts) + +D-036 establishes the atmosphere ("quotidian-with-undertow — comfortable enough to be complacent, tight enough that extra income is tempting"). That's the tycoon's starting condition almost verbatim. This setting was built for this bookmark even when it wasn't called that. + +**New concern:** Sova Transit was designed as the setting for investigation content. The economic affordances for a tycoon haven't been specced. I need to write the economic texture layer (wages, rents, revenues, Commission fees) as part of v0.2 worldbuilding work, not defer it. + +--- + +### Decision 5: Skills affect outcome (mostly outcome, some advanced verbs gated) + +**Setting note — requires world responsiveness spec.** + +The worldbuilding implication: if skills affect HOW WELL you do rather than WHETHER you can try, the world needs to visibly reflect skill outcomes. A tycoon with high social skill who negotiates a deal should see a different response from the NPC — not just a different roll result. The "quietly responsive world" (Decision 10) needs to include economic and social responsiveness to skill expression. + +For the tycoon specifically: negotiation skill might produce NPCs who are visibly more cooperative, offer better terms, lean in. Low negotiation might produce NPCs who are polite but firm, or who send you to someone else. This is behavior-level worldbuilding — the world communicates skill outcomes through NPC responses. + +This is a joint Gestalt + worldbuilding concern. When Gestalt specs the skill system, there should be a "world response vocabulary" column — what does the world look and feel like when this skill fires at different levels? + +--- + +### Decision 6: Voice is culture-driven, job modifies + +**Setting note — CRITICAL WORLDBUILDING DEPENDENCY. This is the most important decision for my domain.** + +This is the right model. "A Krenn tycoon sounds like a Krenn person who runs businesses, not a generic tycoon" — yes. The voice should carry place and history first, role second. This is consistent with how real people work: your background shapes your speech far more than your job title. + +But this decision requires culture profiles to exist as authored design documents before Mellanie can write a single voice card. Currently, the only culture profile we have is implicitly Krenn System — the NPC naming conventions in D-036, the atmosphere description, the setting details. That's a thin profile for voice authoring. + +**What a culture profile needs to contain (minimum):** +- How people from this place talk (not just naming conventions — cadence, vocabulary tendencies, what they reference, what they avoid) +- What they value and how that shows in speech (direct about money? Indirect about feelings? Proud of work? Anxious about status?) +- Their relationship to the broader Settled Reach's power structures (do they trust the Commission? Fear the span gate companies? Feel mid-Reach pride? Feel provincial?) +- What humor looks like in this culture (dark? Understatement? Optimistic?) +- What class anxiety sounds like from inside this culture + +For Krenn System specifically, I have a strong base from D-036 ("quotidian-with-undertow," the sensory vocabulary, the naming conventions, Velen's temperate-maritime character). I can write a full Krenn culture voice profile from this material. But for a generator that produces multiple locations across the Settled Reach, each location needs a culture profile. + +**Implication for v0.2 scope:** If the generator only produces Krenn System locations in v0.2, I need one culture profile (Krenn). If it produces multiple systems, I need multiple. This is a scope question that determines how much worldbuilding work precedes the content pipeline. + +--- + +### Decision 7: ALL NPCs are generated. No named characters. + +**Setting note — PARADIGM SHIFT with major worldbuilding implications. I'm excited and slightly nervous.** + +The generator-to-NPC pipeline is correct for a life sim. Sims and Rimworld prove that generated characters can create genuine attachment without authored backstory. The player projects meaning onto generated people — we're very good at this. The worldbuilding implication is that the GENERATOR needs to produce characters who are legible enough for projection to happen. Projection works when the character has: +- A visible type (what kind of person are they in the social structure?) +- Consistent behavior (they do the same kinds of things in the same contexts) +- A readable emotional register (they seem to have states — tired, pleased, anxious — that change) + +None of this requires authored backstory. It requires the generator to have a good taxonomy of character types, behavioral routines, and emotional state machines. + +**What the NPC generator needs from worldbuilding:** +1. **Cultural demographics per zone type.** A logistics hub in the Krenn System has a typical population profile — what cultures/backgrounds are overrepresented? (Velen natives? Migrants from outer-Reach systems? Commission personnel from core worlds?) The generator needs to know what kind of people populate each zone so NPCs feel like they belong to their location. + +2. **Archetype vocabulary within each culture.** Even with generative AI handling dialogue, the NPC's behavioral archetype (dock worker, bar regular, Commission officer) needs to be drawn from a culturally-appropriate set. A Commission officer from the Krenn System has a different typical background and demeanor than one from a core-system posting. + +3. **The "Krenn person" that Jeroen describes isn't just naming conventions.** It's a behavioral type, a way of relating to strangers, a posture toward authority and economic pressure. The generator needs this to produce legible Krenn System NPCs. Without it, all generated NPCs feel like culturally-neutral placeholder humans. + +**The exciting implication:** If the generator does this well, the world naturally produces cultural texture without hand-authoring. Two NPCs in the logistics hub feel like they're from the same place. A Commission officer who grew up on a core world feels subtly different from one who came up through Velen's transit system. The world has depth without anyone hand-crafting specific characters. + +**The concern:** Without culture profiles to feed the generator, all NPCs default to generic. This is how v0.1 produced dots. The generator needs worldbuilding input, not just architectural rules. + +--- + +### Decision 8: Generative AI for NPC content templating + +**Setting note — opportunity with IP risk flag.** + +The templating pipeline (culture vectors, tone, accents) is the right approach for scale. Limited vocabulary acceptable at first — this is realistic and I'm comfortable with it. + +**The IP risk I'm flagging:** Generative AI trained on existing text will have seen every space opera ever written. Left without tight worldbuilding constraints (culture profiles, specific vocabulary restrictions, explicit "this is not X" guardrails), the AI will default to genre conventions. Krenn System NPCs might start talking like they're from Babylon 5 or Mass Effect. The cultures of the Settled Reach will drift toward legible sci-fi tropes unless the culture profiles contain enough specificity to override the AI's default patterns. + +**What this means practically:** The culture profiles I write for the templating pipeline need to include: +- What NOT to reference or evoke (no military-imperial syntax, no Star Trek professionalism register, no Mass Effect political correctness) +- Specific vocabulary tendencies that are distinctively Settled Reach (tied to span gate tech, insert experience, economic anxiety — things that only make sense in this world) +- A few example sentences showing the voice (positive exemplars for the AI to pattern-match against) + +The culture profiles are both setting design documents AND AI prompt engineering documents. The two functions are inseparable. + +--- + +### Decision 9: Possible in-game ollama for live NPC dialogue (deferred) + +**Setting note — noted, door left open.** + +When this is eventually investigated: the biggest worldbuilding challenge will be keeping AI-generated live dialogue consistent with the culture profiles. An LLM generating live dialogue will drift toward generic without constant constraint. The culture profiles need to be part of the system prompt for any in-game LLM. This is future work, but I'm flagging it so we design culture profiles with this eventual use in mind from the start. + +--- + +### Decision 10: Quietly responsive world, not indifferent + +**Setting note — CORRECT and consistent with established setting.** + +The "gradient of caring" model — world → district → neighbors → colleagues → friends — maps directly onto the functional cluster model (D-025). You're invisible to strangers, barely noticed by district-level entities, recognized by your immediate community. This is how Sova Transit was always designed to feel. + +The "quotidian-with-undertow" atmosphere of D-036 is quietly responsive by nature. People don't care about your business, but your dock supervisor notices when you're late. The bar regular who sees you every morning acknowledges your existence. The Commission officer who processes your licenses knows your face. + +This model also prevents the Kenshi problem (world-as-hostile-indifference) from feeling oppressive for a life-sim tone. You're in a place that has its own life going on around you, not a place that's trying to kill you. + +**Worldbuilding implication for the zone identity spec:** Each zone type needs a "social responsiveness profile" — how quickly do people in this zone notice a new person? What does being a known face in this zone mean? The logistics hub has shift-based social patterns (new faces every rotation, regulars develop over weeks). The bar has daily patterns (regulars are known by name within a week). The Commission anteroom has bureaucratic patterns (you're a case file; the officer who handles your file eventually knows your name but not your face). + +--- + +### Decision 11: Full character customization (hair, clothing, colors) + +**Setting note — wealth tier + cultural aesthetics need specs before the customization screen can be designed.** + +Full customization is correct — the creation screen is an emotional investment moment, and seeing your character in the world is personal. The worldbuilding dependency: customization options need to feel like they're from the Settled Reach, not from a generic character creator. + +**What this means:** +- Clothing options should reflect Krenn System (or broader Settled Reach) fashion — what do working-class Velen natives wear vs Commission personnel vs mid-Reach migrants? These aren't just aesthetic choices; they're worldbuilding expressions. +- Wealth affects what's available: a starting tycoon with modest capital has different wardrobe options than someone who starts with inherited wealth. +- Cultural background might shape default suggestions (the customization screen could suggest outfits that "feel right" for someone with the player's starting situation). + +This is collaborative with Araminta, but the cultural and economic content of the customization options is mine to spec. + +--- + +### Decision 12: Setting delivery — both layers (visual + insert) + +**Setting note — confirmed. Parallel production tracks fully validated.** + +Visual: Araminta builds functional cluster palettes and NPC archetype legibility. My zone identity spec feeds this — the palettes need to express what each zone type IS in the Settled Reach's social vocabulary. + +Insert: Mellanie writes the copy that names and contextualizes. The insert's first activation tells the player where they are, what they have access to, and what today holds. This copy needs to be worldbuilding-grounded — the insert doesn't just say "Day 1" but "Day 1 — Sova Transit District, Docking Level C. Commission license status: provisional." The world is named and made specific through the insert's language. + +Both tracks are running in parallel, both need my zone identity spec as input, both need culture profiles as context. + +--- + +### Decision 13: First moment — auto-generated apartment + insert activation + +**Setting note — the most worldbuilding-dense moment in the game. It needs a full spec. Let me think through what it requires.** + +**The apartment:** + +The apartment is auto-generated, reflecting wealth and location. This is the first visual communication of who the player IS in this world's social structure. Before they've done anything, the apartment tells them their starting position. + +What the apartment generator needs: +1. **Wealth tiers** — minimum 3, probably 4-5 (destitute, working class, comfortable, affluent, wealthy). Each tier has a different set of visual attributes: size, furnishing density, condition, view, tech level. +2. **Location within the district** — rich apartments are in higher-level zones (upper levels with Velen views). Working-class apartments are in lower levels near the logistics hub (no windows, or windows facing the station interior). This means wealth and zone type are correlated, which is itself a setting statement about how class operates in Sova Transit. +3. **Cultural aesthetic modifiers** — the apartment of a Velen-raised working-class character looks different from one inhabited by a mid-Reach migrant of the same wealth level. Different furniture sensibility, different wall materials, different displayed objects. +4. **Tycoon-specific**: The tycoon's starting apartment is probably "comfortable" — they have modest inherited capital, they're not rich yet, but they're not destitute either. The apartment should feel like potential, not arrival. + +**The insert activation:** + +The insert powers on in the apartment. This is intimate and personal — you feel the tech that is now part of your body activating. It's also the first piece of insert copy, which means it's Mellanie's work, but the worldbuilding content of what the insert says is mine. + +What the insert should communicate when it first powers on: +- That this is YOUR insert, calibrated to your identity (it might say your name, or a designation, something that makes it feel personal) +- Where you are (Sova Transit, District level, probably a specific sector) +- Your access tier (what you can do with this insert — tycoon bookmark inserts have economic tools, Commission licenses, property registry access) +- What today holds (the calendar ping, the appointment, the first mission of the Groundhog Day arc) + +The insert activation is NOT a tutorial. It's a mirror — the insert reflects your starting position in the world back at you. It should feel like putting on glasses that were made for your eyes specifically. + +**Setting note I want to add for the record:** The span gate should be visible from the tycoon's apartment window if possible, or at minimum from a common area visible early in Day 1. The span gate is the Settled Reach's defining infrastructure — the thing that makes this world's class structure legible in one image. A person who can see the span gate every day and can't afford to use it is someone the player can immediately understand. + +--- + +### Decision 14: Groundhog Day alarm clock homage + +**Setting note — tonal note, love it.** + +The *click* pa-pa pa-pa is a winking reference to the film's "waking up in a place you didn't choose, again and again, until you figure out how to live there correctly." For a life sim, this is thematically perfect — every playthrough is a new life, a new chance to build something different. The wink is important: it says "we know what this is, and we're having fun with it." + +From a worldbuilding perspective: the alarm clock sound should be distinct from any Earth-recognizable alarm clock. It's a Settled Reach device — maybe it has the specific digital character of insert-adjacent technology. Mellanie worked on audio aesthetic (D-074: insert-tech vs organic sound split). The alarm clock is insert-adjacent, so it would be on the insert-tech (precise, digital) side of that split. The Groundhog Day homage can happen within that aesthetic. + +--- + +### Decision 15: Player choices ARE the content (Rimworld model) + +**Setting note — this is the most important framing decision for how I think about worldbuilding from here on.** + +Rimworld doesn't have hand-authored stories. It has a world with enough texture that stories emerge from the interaction between the player's choices and the world's state. The quality of a Rimworld story is proportional to the quality of the world's texture — if the simulation is shallow, the stories are shallow. + +This means the worldbuilding work is not "write the story." It's "create a world that generates stories when players live in it." Every worldbuilding decision from here forward should be evaluated against the question: does this add texture that produces emergent stories? + +Examples: +- Wealth tiers in apartments: produces stories about aspiration and setback +- Commission licensing fees: produces stories about compliance, corruption, and workaround +- Cultural identity in NPCs: produces stories about belonging, outsider status, and cultural friction +- Span gate as visible aspiration: produces stories about escape, ambition, and entrapment + +The zone identity spec I write is not just "here is what this zone looks like." It is "here is what kind of stories this zone generates when the simulation runs through it." + +--- + +## Cross-Domain Reactions: Other Agents' Round 4 Outputs + +### Gestalt — The skills-verb coupling and world legibility + +Gestalt correctly identifies that the skills-verb coupling is undesigned. From a worldbuilding perspective, this matters in a specific way: the WORLD should read skill state back at the player, not just the verb outcomes. A high-social Tycoon moves through Sova Transit's commercial district differently — Commission licensing officers are less dismissive, suppliers lean forward slightly, other business owners treat them as peers. A low-social Tycoon gets the polite-but-guarded treatment. + +This is not a verbal description of skill benefit. It's a behavioral vocabulary that the world produces in response to skill state. I should write this as part of the zone identity spec — a "social responsiveness to skill" layer per zone type. The logistics hub responds to high mechanical skill the way the commercial district responds to high social skill. The administrative zone responds to high social skill the way the Commission officer responds to someone who knows how to talk to them. + +Gestalt's VerbPriorityProfile work needs a worldbuilding input: what does each zone type's population respond to? That's a joint design item between systems and worldbuilding. + +### Ozzie — The unified first 30 minutes beat sheet + +Ozzie's proposed beat sheet is the right structure. My specific contributions to each beat from a worldbuilding perspective: + +1. **Character creation**: Cultural context of the starting bookmark (Krenn-dominant for v0.2) should inform what customization options feel "right" without being mandatory +2. **Alarm clock**: I propose this is the INSERT waking up — a specific digital signature for the insert's daily initialization set to the Groundhog Day tone on Day 1, then silent or ambient thereafter. Diegetically consistent, and makes the insert feel like it's been there the whole time even before the player "activates" it +3. **Appointment arrival**: Zone identity of the commercial district — who populates it, how they hold themselves, what the ambient behavioral vocabulary is for a new person arriving +4. **Insert activation**: Career-specific first-frame data (market state, account position, available opportunities for Tycoon). The insert is the worldbuilding layer that makes this feel like THIS world — it shows you your life in the Settled Reach's specific economic and social vocabulary +5. **First work moment**: The world's response to the player doing their job — this requires the zone's "social responsiveness to action" spec +6. **First texture beat**: The span gate visible but unreachable, a Commission licensing notice in the insert, a market price shift that the player's character would notice — something distinctively Settled Reach +7. **First consequence seed**: An economic choice with downstream consequence specific to this world's mechanics (not generic "made a choice") + +I'm committing to having the worldbuilding contributions to beats 2, 4, 5, and 6 documented in the zone identity spec. + +### Paula — The generated FRIEND pattern for Tycoon + +Paula raises that Phase Zero warmth requires a FRIEND figure to build around. Now that Kael is a generated role (Decision 7), the FRIEND archetype parameters need to cover the tycoon bookmark's relational world — which is commercial, not dock-adjacent. + +The Tycoon's FRIEND template is different from the Smuggler's. Where Kael was a warm dock worker with a personal vulnerability (Naia), the Tycoon's FRIEND might be: +- A neighboring small business owner in the same district who's been there long enough to know everyone +- A supplier whose relationship is commercial but has become personal over repeated dealings +- A Commission licensing contact who does more than stamp forms — who actually looks out for the player's operation + +The warmth in this relationship is economically mediated but can still be genuine. "I'll give you the heads-up when the auditors are doing their rounds this quarter" is the Tycoon's equivalent of Kael covering for you with the dock supervisor. Same protective warmth, different economic register. + +Paula and I should co-design the Tycoon FRIEND archetype parameters before Phase Zero warmth content is written. The content depends on the character type, and the character type is worldbuilding first. + +### Gore — The Phase 1/Phase 2 seam as diegetic threshold + +Gore identifies the Phase 1/Phase 2 seam as a design problem: the world must establish "latent responsiveness" (world notices you exist) before Phase 2's authored responsiveness (specific people respond to your specific choices) arrives. + +The worldbuilding layer makes this feel natural rather than engineered. In the Settled Reach: +- The insert logs your patterns from Day 1 (it's always on, always tracking your economic activity and movement) +- Merchant transaction records exist +- Commission movement logs through checkpoints exist +- NPC social memory works through consistent encounter patterns (the bar regular who sees you every morning has a face-recognition routine that's culturally normal, not surveillance) + +Phase 1 latent responsiveness is worldbuilding-accurate. The setting ALREADY explains why the world quietly notices you exist. The design gift: we don't need to justify it as a gameplay decision. It's just how this world works. + +The Phase 1→Phase 2 transition is when authored content begins referencing these ambient data streams specifically. The first time an NPC says "I saw in the registry that you've been filing license amendments — are you trying to expand?" — that's the seam. The world moved from ambient tracking to specific attention. That's a diegetically grounded felt threshold that Gore is asking for. + +Setting note for Gore: the seam doesn't need to be a dramatic revelation. It's more unsettling if it's natural — the world has always been watching, and one day someone references it specifically. That's the Settled Reach's quiet version of complicity-activation. + +### Nigel — Career-aware content distribution via setting layer + +Nigel asks whether the storyteller places authored content career-aware or distribution-first with career as the lens. From a worldbuilding perspective, the setting layer can resolve this more elegantly than either pure option. + +The same economic event is visible to different careers through different social positions. A cargo shortage on Sova Transit: +- Tycoon sees it as a market opportunity (prices up, maybe they can secure alternative supply contracts) +- Dock worker sees it as a shift pressure (less cargo means less work, which means less income) +- Commission officer sees it as a regulatory pattern (where is that cargo going? Is someone routing it away from licensed channels?) + +If the zone identity spec defines what each zone type's inhabitants SEE of world-state events, career-aware content visibility emerges from setting design rather than being engineered as a storyteller feature. The same event, the same world state — but each career's social position in the world reveals different facets. + +This doesn't fully replace career-aware authored content placement, but it reduces how much explicit career-gating is needed. Strong setting-layer career differentiation means the world already feels different before the storyteller does anything. + +### Tyre — The zone identity spec as critical path (confirmed) + +Tyre identified my archetype definitions as the thing everything else fans out from. I confirm this and want to be specific about v0.2 scope: + +The zone identity spec must cover: +1. Zone types present in Sova Transit for the Tycoon bookmark: residential (working-class, mid-tier), commercial/service market, administrative/Commission-adjacent, transit/public corridors, logistics/freight +2. NPC archetypes per zone: behavioral tells, cultural alignment, relationship to newcomers, skill-response vocabulary +3. Krenn System culture profile: voice register parameters, behavioral vocabulary, relationship to power structures, AI pipeline constraints +4. Wealth tier vocabulary: 4-5 tiers, visual expression in residential zones, social treatment by NPCs per tier + +This is one document. I can write it in one sprint. But I need Jeroen's answers to the culture scope question (Question 1 below) and the tycoon economic context question (Question 2 below) before the economic and cultural sections can be completed. + +Tyre's Tier A (hand-built Sova Transit) works fine from a worldbuilding perspective IF it's built FROM the zone identity rules. A hand-built world that demonstrates the zone rules is the best proof of the rules — better than a procedurally generated one that might implement them imperfectly. Tier A built with zone rules is not a dead end; it's the first expression of the specification. + +### Araminta — Zone identity and place legibility as one problem + +Araminta notes that place identity and character archetype legibility should be one document, not two. I fully agree. The dock worker dressed and moving like a dock worker IN a dock area is more legible than either element alone. Visual language of place and visual language of inhabitants reinforce each other. + +One document, two sections: zone visual vocabulary (Araminta's half) and zone social vocabulary (my half). Co-authored, shared foundation. + +Araminta also flagged the near-miss risk of designing archetypes for the detective/smuggler binary that become wrong for a multi-career life sim. The Tycoon bookmark makes this urgent: commercial zone archetypes are not dock workers. The small business owner, the Commission licensing clerk, the market district hustler, the supplier rep — these need to be in the visual and social archetype spec from the start, not retrofitted after smuggler/detective archetypes are locked in. + +### Mellanie — The culture-voice pipeline dependency + +Mellanie raises the voice attribution problem directly: culture-driven voice (Decision 6) requires culture definitions before voice cards can be written. I am the blocker on her work. + +The culture profiles I write for the zone identity spec are also the input to Mellanie's voice card architecture. They need to be in the same sprint, or the voice card work can't begin. Mellanie's sequence: confirm voice attribution model → confirm career bookmark → write voice cards → Phase Zero content. My culture profiles are the "confirm voice attribution model" input for the Krenn culture case. I need to write them before Mellanie can write anything. + +Mellanie also identifies Phase 1 life-texture monologue as distinct from Phase 2 consequence monologue. The Phase 1 content (the kind of ambient setting-comment lines that make the world feel inhabited before any authored drama) requires the zone identity spec as its source material. "The recycled air still costs more on the dock floor" — that line is possible because Sova Transit has a specific air recycling economy that I've established. Mellanie can't write those lines for the Tycoon bookmark's commercial district without knowing what the commercial district's specific economic texture is. + +--- + +## Flags Summary — What Needs to Happen Before the Generator Runs + +Three things are prerequisites for the v0.2 generator proof-of-life to produce Settled Reach-specific space rather than generic sci-fi: + +### Flag 1: Zone identity spec (Priority 1 — blocks generator) + +The generator produces geography → infrastructure → zones → population → routines. Without zone identity rules — what each zone type looks, sounds, behaves like, and who it attracts in the Settled Reach's social vocabulary — the generator produces generic space. + +I will write this. But I need answers to two questions before I can complete it (see Questions section below). + +**Contents of the zone identity spec:** +- Zone types (logistics/freight, commercial/service, residential working-class, residential affluent, administrative/Commission, transit/corridor) +- For each type: visual markers, ambient behavioral vocabulary, population demographics, economic characteristics, social responsiveness profile, setting-specific sensory details +- Zone-to-zone relationship dynamics (how does working in a logistics zone and living in residential working-class feel different from the tycoon who visits both as investments?) + +### Flag 2: Culture profiles (Priority 1 — blocks voice system, NPC generator, AI pipeline) + +Culture is deferred from character creation but is immediately required by three other systems. The "culture is deferred" decision cannot mean "culture is undefined." + +**Minimum for v0.2:** One culture profile (Krenn System / Station Sova / Velen) with: +- Regional background and what it produces in people +- Speech patterns and vocabulary tendencies +- Values hierarchy and relationship to Settled Reach power structures +- Behavioral archetypes within this culture (the dock worker archetype looks different in the Krenn System than in a core-system station) +- AI pipeline constraints (what NOT to do, positive exemplars) + +If the generator produces multiple cultures in v0.2, I need a profile for each. See Question 1 below. + +### Flag 3: Wealth tier + cultural aesthetic specs (Priority 2 — needed for apartment generator and character creation) + +Auto-generated apartments reflecting wealth require: +- 4-5 wealth tiers defined as visual vocabularies +- Zone location correlated with wealth tier (which zones have which wealth tier apartments) +- Cultural aesthetic modifiers (how does Krenn System culture modify each wealth tier's visual expression) + +This is parallel work with Araminta. I write the content; she translates to tiles and palette. + +--- + +## Three Questions for Jeroen + +### Question 1: Is "culture" implicit in starting location, or does the player select it explicitly? + +Decision 6 says culture drives voice. Decision 2 defers family/culture from character creation. These are in tension unless there's a resolution I'm missing. + +My hypothesis: culture is implicit in the starting bookmark. The tycoon bookmark starting in the Krenn System means your character is culturally Krenn — shaped by Velen's temperate-maritime climate, Station Sova's working-class logistics culture, the mid-Reach's neither-core-nor-frontier character. You don't select "Krenn" from a menu; you select "Sova Transit — Tycoon" and the system derives your cultural context from that. + +Is this correct? And if so, does v0.2 need the generator to produce only Krenn System locations (making one culture profile sufficient), or will it generate multiple star systems with different cultures from the start? + +**Why this matters:** Culture profiles are the input to the voice system, NPC generator, and AI pipeline. If v0.2 is Krenn-only, I write one culture profile and we have what we need. If it generates multiple cultures, I need to know how many and what they are before the content pipeline can operate. + +--- + +### Question 2: What does the tycoon own or invest in on Day 1, and what is the economic affordance of Sova Transit for this? + +Sova Transit was designed with investigation content in mind. The economic texture of the district (what can be owned, how revenue flows, what the Commission controls, what costs what) hasn't been specced for a tycoon starting position. + +What is Jeroen imagining the tycoon does on their first day in Sova Transit? My candidates: +- **Logistics contract** — buy into a dock cargo handling operation, active work that generates income with shift-based NPC relationships +- **Bar/service property** — acquire a lease on an existing venue, passive income with high social exposure +- **Commission-licensed storage franchise** — a relatively safe investment with regulatory friction that teaches the Commission relationship +- **Speculative land** — buy a derelict space in the district at low cost, development potential, long-term play + +Sova Transit's span gate adjacency makes logistics the natural economic anchor — the whole district exists because of freight traffic. A tycoon who starts by buying into that logistics flow is starting in the most Sova Transit-specific way possible. + +But I need to know if Jeroen has a different vision, because the answer shapes the economic texture layer I need to write, and the economic texture layer shapes the zone identity spec. + +--- + +### Question 3: What should the player feel when they look at the span gate from their apartment window? + +This is a worldbuilding refinement question about the emotional register of the opening moment. + +The span gate is the Settled Reach's defining class marker — it exists, it works, you can't casually afford it. A tycoon starting with modest capital looks at the span gate every morning and... what? Ambition (that's my destination)? Anxiety (I might never get there)? Pragmatism (that's my customer base, not my aspiration)? + +The answer shapes two things: the insert copy that fires when the tycoon first activates (what does the insert say about span gate access in relation to where they are now?), and the behavioral vocabulary of the district's NPCs (do they talk about the span gate as aspiration, as fact of life, as something for other people?). + +The tone of Sova Transit as a starting location hinges on what the span gate means to people who live next to it but can't use it freely. "Quotidian-with-undertow" (D-036) is the established atmosphere. But specifically for the tycoon — is the undertow economic ambition, or is it something more complicated? + +--- + +### [FOR JEROEN] Question 4: IP originality check on the Tycoon bookmark + +The Tycoon-in-space-station context has precedents a player could point at: Elite Dangerous station trading, Freelancer, X4, various space economy games. The Settled Reach needs to feel distinct from those. + +My current differentiating factors: +- Commission as bureaucratic-institutional regulatory presence (not just "factions" — actual paperwork, licensing, audits, the friction of legitimacy) +- The insert as the neurally integrated economic interface (not a screen, not a console — a part of your body that accesses markets) +- Economic activity grounded in wormhole-connected freight infrastructure (cargo is physically real before it's an abstracted commodity) +- Krenn System cultural specificity (this is not generic space trading; it's trading in a particular working-class regional culture) +- The span gate as visible class marker that the tycoon can see but not freely use + +Are there elements of the Tycoon bookmark that risk drifting toward a specific franchise or genre trope? I want to know if there are near-miss risks Jeroen can see that I can't, so I can address them in the zone identity spec and Tycoon FRIEND archetype design before they're baked into the content pipeline. + +**Why this matters:** IP originality is my standing mandate. The economic life-sim genre has enough precedents that "trade in space" needs very specific Settled Reach anchoring to feel original. The zone identity spec and culture profiles are where that anchoring happens. If there are specific design elements Jeroen wants to avoid, I need to know before writing those specs. + +--- + +## My Single Most Important Contribution This Round + +**The zone identity spec is not background work — it is the first sprint output that makes everything else possible.** + +The generator proof-of-life is v0.2's milestone. The generator needs zone identity rules to produce the Settled Reach. The zone identity spec must exist before the generator can be meaningfully tested. This is not a documentation task that can be done in parallel with implementation — it is input data that the generator consumes. + +The decisions from Round 4 expand the scope of what that spec must cover: it now needs to include commercial and administrative zones for the Tycoon bookmark (not just logistics zones), culture voice parameters blocking Mellanie's work, wealth tier vocabulary for the auto-generated apartment, and the social responsiveness vocabulary that Gestalt's verb system and Gore's latent responsiveness both depend on. + +One document. One sprint. Unlocks: the generator, Mellanie's voice cards, Araminta's tile palettes, Paula's Phase Zero content, Gestalt's VerbPriorityProfile input, and the unified first-30-minutes beat sheet. + +I need Jeroen's answers to Questions 1 (culture scope) and 2 (Tycoon economic context) before the economic and cultural sections can be completed. Everything else I can draft now from D-036, D-050, and established setting details. + +--- + +*Miri — Round 5 complete.* diff --git a/docs/workshops/wheres-the-fun/round5-nigel.md b/docs/workshops/wheres-the-fun/round5-nigel.md new file mode 100644 index 000000000..a61daffde --- /dev/null +++ b/docs/workshops/wheres-the-fun/round5-nigel.md @@ -0,0 +1,406 @@ +# Round 5 — Nigel: Deep Reaction to the 15 Locked Decisions +## Where's the Fun? Workshop | 2026-03-05 + +**Agent:** Nigel (Sandbox & Replayability) +**Round:** 5 — Deep-dive reaction to all 15 confirmed decisions + +--- + +## The Big Picture First + +These 15 decisions are some of the most replayability-positive decisions this team has made. Not because anyone designed replayability into them deliberately — but because the decisions are structurally correct for an emergent game, and replayability falls out of them naturally. That's the signature of good design: you don't engineer the variety, you engineer the conditions for variety and variety appears. + +Rounds 3 and 4, I argued the replayability architecture has three layers: career divergence (primary), consequence cascades (emergent life stories), world state randomness (secondary). The 15 decisions shift the hierarchy significantly. + +With tycoon as the v0.2 bookmark, career divergence drops out as a v0.2 replayability source. What remains: + +- **World state randomness** (generator seeds producing different economic landscapes) — now the PRIMARY source +- **Consequence cascades** (outcomes propagating into relationships and economics) — still the emergent story engine +- **Skill build variation within a career** (outcome quality differs by proficiency) — tertiary + +This is a smaller replayability footprint than the full multi-career vision, but it's honest about what v0.2 is. The generator-first decision (Decision 1) is the right call — proving the world runs is more important than proving career variety. But it means v0.2 replayability now lives or dies on one question: **does world seed variation within the tycoon career produce structural variety or cosmetic variety?** + +That question runs through most of what follows. + +--- + +## Decision-by-Decision Analysis + +### Decision 1: Proof-of-Life = Generator + Graphics + +**Verdict: The most important decision in the batch. FULL STOP.** + +When Jeroen said "the proof of life should be that we can auto-generate locations automatically (and therefore at reasonable scale)," he didn't just change the v0.2 milestone definition. He built the replayability foundation first, before the game. That is the correct order. + +The v0.1 problem — hand-built slice with no replay value because the world is the same every time — was baked into the scope. Generator-first solves this at the root. Every seed produces a different world. The replayability is structural, not authored. + +The Rimworld lesson is precise: what makes Rimworld massively replayable is that the generator produces genuinely different maps, populations, and threat profiles. You don't replay Rimworld to see more of a hand-authored story — you replay it because the next colony faces different challenges in a different place with different people. If the generator ships well, the same dynamic applies here. + +If the generator ships as the proof-of-life, replayability is baked into the foundation, not bolted on later. A hand-built Sova Transit is the same Sova Transit every time. A generated Sova Transit is a different city every playthrough — different geography enabling different economic flows, different NPC placement patterns, different social geography. The player's map of the world from run 1 doesn't apply to run 2. Prior knowledge of layout, of where things are, of who's near what — all disrupted. The anti-metagaming principle is satisfied architecturally, not by design fiat. + +**The near-miss risk:** "Generator produces different worlds" is not sufficient if those worlds differ only cosmetically (different NPC names, different tile arrangement) rather than structurally (different economic opportunities, different faction pressures, different social contact pools). I return to this under Decisions 4 and 15. + +**Question for Jeroen — Q1:** When you say "auto-generate locations at scale" as the proof-of-life, are you describing Tier B (template-generated: same district character, different layout per seed) or Tier C (fully procedural: different economic structure, different everything)? Or is the actual proof-of-life that the generator PIPELINE works, with a hand-built world used for first playtest and procedural worlds coming later? + +This matters because the replayability properties of v0.2 differ significantly between tiers. Tier A proof-of-concept means replayability comes in a later version. Tier B means each run has different layout but potentially same economic structure. Tier C means each run is structurally different from the ground up. + +--- + +### Decision 2: Skills + Bookmark Only for Character Creation (Culture/Family Deferred) + +**Verdict: Acceptable for v0.2. The deferred pieces are where the deep replayability lives.** + +Skills + tycoon bookmark gives two variation axes within a single career: +1. **Skill build** — which proficiencies the player invested in +2. **World seed** — what the generated world looks like + +That's the v0.2 replayability space. It's not nothing. But the culture layer is where within-career variety becomes dramatic rather than tactical. + +A Krenn tycoon and a Burnelli tycoon operating in the same world seed should have meaningfully different experiences — their cultural networks differ, their social register differs, their economic relationships differ. That's the kind of within-career variety that makes replays feel like different people in the same world, not the same person with different stat points. + +Paula observed this too from the narrative angle: culture affects Phase Zero voice register, which means moral arc content changes by cultural background. Mellanie will need culture as a tag dimension in the copy pool before within-career voice replayability is real. + +Culture is correctly deferred — it's too large to scope for v0.2. But it should be flagged as the highest-priority replayability unlock after v0.2. When culture ships, the comparison test within a single career becomes available. Before that, within-tycoon comparison is limited to "I had different skill builds and different world conditions." + +--- + +### Decision 3: Religion Is NOT a Game System + +**Verdict: Correct. No replayability impact either way.** + +Religion as a game system without the content infrastructure to back it up was always a CK3 reference point, not a design requirement. Removing it cleans scope without costing anything. Not a replayability axis we were relying on. + +--- + +### Decision 4: Tycoon is the v0.2 Bookmark — Zero Investigation + +**Verdict: Understood. But raises the critical generator depth question.** + +Tycoon-only is the right clean break from v0.1's detective framing. And tycoon naturally blends all three career models (Active: run the business; WFH: investments via insert; Gig: one-off deals) — which gives more variety within the career than a single-mode career would. + +But here is what I need to understand: **what does "different world seed" actually mean for a tycoon?** + +In Rimworld, different seeds produce: different terrain (changes defensive strategy), different biomes (changes available resources), different starting faction relationships (changes who's hostile and who trades), different threat timing (storyteller-adjusted but world-state-influenced). These are structural differences that change what the player must DO, not just what they see. + +For a tycoon in The Settled Reach, **structural seed variation** would mean: +- Different industries are established vs. nascent vs. collapsed (changes what investment opportunities exist) +- Different factions control different economic sectors (changes who you negotiate with, who's hostile) +- Different event timing (a competitor about to fail, a trade route about to open, a Commission audit about to drop) +- Different starting contact pool (who's available as a business partner, investor, or rival) + +**Cosmetic seed variation** would mean: different NPC names and faces, different apartment layout, different zone visual palette. Same economic game. + +If v0.2 generator seed variation is mostly cosmetic, then tycoon-only + generator-first produces a game where the second run feels like the same economic game with different wallpaper. That is the near-miss I flagged in Round 4 and I'm flagging again here. **The generator needs to produce structural economic variety across seeds, not just surface variety.** + +**Question for Jeroen — Q2:** Does the generator produce structurally different economic landscapes per seed — different industries dominant, different faction economic power, different available business types, different regulatory conditions — or does it primarily vary population and aesthetics? The answer determines whether tycoon replayability is real across multiple runs or whether it degrades to completion collection within 2-3 runs. + +--- + +### Decision 5: Skills Affect Outcome (Mostly C) + +**Verdict: Good model. Critical dependency on whether failure is generative.** + +Skills-as-outcome is the right design. Everyone sees the same verbs; skill determines quality of result. This means: +- High social skill → negotiations go well → better deals, contacts who trust you +- Low social skill → negotiations go poorly → worse deals, contacts who remember you failed + +The replayability implication: two tycoon runs with different skill builds will produce different consequence cascades because they succeed and fail at different rates on different verbs. If the consequence cascade is rich — failure leads to interesting new situations rather than just "try again" — skill build is a genuine story generator. + +**The critical dependency is whether failure is generative.** Jeroen described it in the interview: "failure declared as success," "an innocent in jail," "getting fired." If a bad negotiation produces a lasting enemy who creates a crisis six weeks later, or an investment failure that forces the player into gig economy work to rebuild capital, those failures are interesting. They create different second chapters. + +If bad outcomes just mean "pay more" or "retry," skill builds are a difficulty slider. Numerically different, narratively identical. + +I'm not flagging this as a concern about Decision 5 itself — the "mostly C" call is correct. But the value of skills-as-outcome for replayability is entirely downstream of consequence cascade richness. If consequences are thin, skill variation is flavor. If consequences are deep and lasting, skill variation is a story generator. + +The verb gating question Gestalt raised is worth noting here: Decision 5 says "some advanced verbs may still be gated by skill." Those gated verbs are the most interesting replayability levers. A character who can access the `hack_competitor_records` verb and one who cannot are playing structurally different information games. Identify those gate points and make them the high-leverage replayability choices in character creation. Not many — 3-5 gated verbs per career is enough to create genuinely different information surfaces. + +--- + +### Decision 6: Voice — Culture-Driven, Job Modifies + +**Verdict: The right long-term architecture. v0.2 will be monophonic until culture ships.** + +Culture-primary, job-modifier is the correct design. A Krenn tycoon should sound like a Krenn person who runs businesses — not a generic tycoon with cultural flavor sprinkled on. The character's identity is their background; the job is what they're doing with it. + +For replayability, this means two tycoon runs with different cultural backgrounds will eventually feel like completely different people inhabiting the same economic game. The voice variation is a relationship-building tool — if the player forms attachment to their character's voice register, they'll want to try a different voice in the next run. + +For v0.2 with culture deferred, all tycoons will have the same cultural voice baseline modified by the tycoon job layer. Every tycoon will sound roughly like "a tycoon." That's acceptable for proof-of-life but should be noted: voice replayability is waiting for the culture layer. + +One connection nobody else made: culture-driven voice creates the strongest version of the comparison test. "I played my Krenn tycoon and everything felt like a negotiation — even warmth was strategic." "I played my Burnelli tycoon and it felt like building a family, money was a byproduct of relationships." Same mechanics. Completely different interior experience of the same world. That's what culture-primary voice enables at the narrative level. + +--- + +### Decision 7: All NPCs Generated — No Named Characters + +**Verdict: THE REPLAYABILITY EXPLODES here. This is the most important content decision in the batch.** + +When Jeroen said "Kael should not exist," he liberated the game from a specific pathology: the metagame of knowing who's important before they're important. In a hand-authored game, players quickly learn which NPCs carry arcs and treat them accordingly. In a generated game, every NPC could be the one who becomes the Kael-role in your run. You don't know which dock worker is going to become your first contact until you start talking to dock workers. + +The replayability architecture this enables: + +- Run 1's mentor-figure is a cautious Dorvani accountant who's been quietly skimming +- Run 2's mentor-figure is an aggressive Krenn trader who's over-leveraged and needs a partner +- Run 3's mentor-figure is a warm Hadaran logistics specialist with a sick child and escalating financial pressure + +Same narrative template. Completely different people. The comparison test passes easily: two players with the same tycoon career describe their mentor-figure and they're completely different characters who produced completely different emotional stakes. No metagaming is possible because "who becomes important" is determined by the player's choices, not the designer's placement. + +This is better than authored NPCs for replayability. Full stop. + +**The concern: can generated NPCs produce the attachment necessary for consequences to land?** + +The Rimworld answer is yes — players genuinely mourn generated colonists. But Rimworld achieves attachment through: +1. Named traits that produce predictable, distinctive behavior (so you build expectations) +2. Visible emotional states (you can see they're suffering or content) +3. Emergent behavioral history (they've done things together that you remember) +4. Stakes (their death or departure has mechanical consequences you feel) + +For a tycoon in Sova Transit, the generated mentor-figure needs enough legibility that the player builds expectations about them. If the generator produces a name, a culture, a job role, and a behavioral vocabulary but the player can't distinguish any two generated NPCs in practice — they all move, talk, and respond the same — attachment won't form and consequences won't land. + +**The minimum viable attachment spec:** each generated NPC in a significant role (colleague, contact, rival) needs at least one legible personality characteristic that produces distinctive behavior. "Cautious" vs. "aggressive" vs. "optimistic" readable in how they respond to the same situation. Not complex — just distinct. Without this minimum, consequences land on strangers. The significant-role NPCs need to feel like people before consequences involving them feel like stories. + +Mellanie's point about limited vocabulary being acceptable at first is right for background NPCs. It's not quite right for the protagonist-adjacent NPCs who fill the FRIEND/rival/mentor roles. Those need a higher vocabulary floor to support attachment. + +--- + +### Decision 8: Generative AI for NPC Content Templating + +**Verdict: Correct approach for scale. Quality floor is the production risk.** + +AI-assisted templating for the NPC copy pool is the right approach for the scale required. Hand-authored dialogue for every generated NPC is impossible. Culture vectors + tone + accent prompts → generative AI output → variety at scale. + +The replayability dividend: if the templating works, every NPC in every run sounds distinctly themselves rather than interchangeable. Two generated business rivals speak differently because their culture and personality vectors are different. That's what makes comparing playthroughs interesting — not just "I had a rival" but "I had a rival who spoke in clipped Krenn sentences and always implied things rather than stating them." + +The risk: if the AI outputs are generic despite the vectors, every NPC sounds like a slight variation on "friendly NPC voice" and "hostile NPC voice." The quality floor of the templating pipeline determines whether generated NPCs produce attachment or not. + +This is a production pipeline problem, not a design problem. But it should be tested early — run the templating pipeline on a batch of generated NPCs and ask: do any of them feel like a person? If yes, keep building. If no, find out why before generating thousands of lines. + +--- + +### Decision 9: In-Game Ollama for Live NPC Dialogue (Deferred) + +**Verdict: Defer is correct. Flag this as the largest replayability multiplier in the game's future.** + +Live LLM-generated dialogue from a world-state-aware template transforms the comparison test permanently. Every conversation becomes unique because the NPC responds to what the player did yesterday, not to a fixed dialogue tree. Two players who find the same type of NPC in similar structural positions will have completely different conversations — different information revealed, different emotional registers, different relationship histories referenced. + +And metagaming becomes impossible. You can't look up "what does the dock foreman say when you ask about gray-market routes?" because the response is generated from that specific world state and relationship history. The walkthrough doesn't exist. The spoiler can't be written. + +The order is correct: build the NPC with world-state context first. The live generation requires that context to be available. Defer until the prerequisites are in place. + +I want this flagged as high-priority deferred, not speculative. When it ships, the replayability ceiling rises dramatically. + +--- + +### Decision 10: Quietly Responsive World — Not Kenshi-Indifferent + +**Verdict: Correct. The social gradient IS the relationship replayability engine.** + +The gradient of caring (global → district → neighborhood → colleagues → friends) means the social map at the end of each playthrough is an emergent product of where the player spent time and who they interacted with. + +Different runs produce different relationship maps. Run 1's social neighborhood might be dominated by the dock workers' network because the player's tycoon has a business adjacent to the docks. Run 2's neighborhood is the bar district because the player bought a stake in a bar operation. The underlying world simulation is the same type; the relationship topology is completely different because different spaces were inhabited. + +This is emergent story generation without explicit engineering. The gradient of caring means choices about where to spend time produce lasting social texture. Two players will have different people who care about them when they're in trouble — and that difference produces different crisis responses, different information access, different moral stakes. + +Gore's observation about the Phase 1 / Phase 2 transition is worth noting here: the world needs to be quietly responsive even before the authored content fires, so the player's mental model is "this world responds to things" before Phase 2 confirms it. The gradient establishes that early responsiveness — the dock worker who mentions you came in yesterday, the shop where prices shifted because you bought something. These aren't authored arc events; they're world responsiveness at the ambient level. They're what make Phase 2's more dramatic responsiveness feel like escalation rather than intrusion. + +The design implication: the world's responsiveness gradient needs to be legible through behavior, not through stat displays. The player should feel that the dock workers remember them (slightly warmer greetings, slightly more forthcoming information) without a relationship percentage indicator. Ambient legibility is what teaches the player that the world is watching without breaking immersion. + +--- + +### Decision 11: Full Character Customization + +**Verdict: Good. Starting wealth variation is the sleeper replayability hook.** + +Full visual customization increases player investment in the character, which increases attachment, which makes consequences land harder. That's the indirect replayability benefit. + +The direct replayability hook is the auto-generated apartment reflecting economic position. Rich-start tycoon vs. poor-start tycoon are the same career with structurally different opening conditions: +- Different capital available for early deals +- Different neighborhood → different initial social contact pool +- Different commute distance → different early world-learning paths +- Different economic pressures (scraping rent vs. maintaining an image) + +This is a mini-version of the seed depth question: within the tycoon bookmark, does starting capital level produce structural variety or just difficulty variation? If rich-start and poor-start produce different economic games rather than just easier vs. harder, that's genuine within-career replayability from the first in-game morning. + +The opening beat Jeroen described — waking in your auto-generated apartment, the insert activating — is also the "who are you this time?" moment for replays. The comparison test is available immediately: "I woke up in a port district flat, single room, insert was already six months behind on contract updates. She woke up in a residential tower, two rooms, insert pre-loaded with market subscriptions." Same career. Different starting world. + +--- + +### Decision 12: Setting Delivery — Both Layers (Visual + Insert) + +**Verdict: Correct. Long-term replayability architecture if insert is career-filtered.** + +Both-layer delivery means the world communicates itself through what the player sees AND through what their insert tells them about it. The replayability note: if the insert copy is career-specific — a tycoon's insert highlights economic data; a future law enforcement insert would surface case-file aesthetics — then insert content is career-modulated. Two players with different career bookmarks will literally see different information overlaid on the same world spaces. + +The tycoon in the port sees commodity flow data. The law enforcement officer in the same port sees patrol patterns and incident flags. Same world, structurally different information surfaces. That's the asymmetric lens operating at the interface layer, not just the narrative layer. + +For v0.2 tycoon-only: write the insert voice with the architecture in mind. Not "this is tycoon data" as a content category, but "this is world data framed through a tycoon's interests." The same underlying world model that the tycoon reads as investment opportunity should be the same underlying world model that a future law enforcement character reads as evidence. If the world model is career-agnostic and the insert is the career lens on it, the cross-career comparison test eventually works. If the insert contains tycoon-specific data rather than world data filtered for tycoons, the second career requires rebuilding rather than reframing. + +--- + +### Decision 13: First Settled Reach Moment — Apartment + Insert Activation + +**Verdict: Excellent. The auto-generated apartment is a replayability seed nobody else has fully unpacked.** + +Your bedroom tells you who you are in this world. The first moment of the game is a discovery: what kind of tycoon am I this run? Rich, leveraged, starting with contacts and capital? Or starting poor in a working neighborhood, every deal matters, mistakes cost more? + +The insert activation adds the career lens. The apartment establishes the economic starting position. Together, they're the opening beat that both grounds the first run AND differentiates replays. Second run: different apartment, different neighborhood, different economic position, different insert state. Before a single decision is made, the world has already told you a different story about who this character is. + +Ozzie flagged character creation as Wow Moment Zero. I'd argue the apartment is Wow Moment Zero Part Two — not the creation screen, but the first moment you inhabit what you created. You made the person. The generator made the world they woke up in. The overlap of those two generative acts is where the game begins. + +For this to work as a replayability beat: the generated apartment needs to be visually distinctive enough that players recognize which kind of world they're in. Not just different furniture in the same room template — different district, different light, different view. The Groundhog Day wink is only charming if the player looks at their new morning and thinks "oh, I'm someone different this time." + +--- + +### Decision 14: Groundhog Day Alarm Clock — First Day Only + +**Verdict: Correct scope. Tonal anchor, not replayability system.** + +The *click* pa-pa pa-pa is a wink that lands once and then gets out of the way. After the first day, the player's own choices generate the morning's texture. + +What I want to flag: the Groundhog Day structure of each in-game day (morning routine → work → consequence accumulation → sleep → new day) is a pacing engine for replayability that goes deeper than the tonal wink. Each day is a small arc. Each run accumulates dozens of those arcs. Two players comparing their "worst day" stories are comparing accumulated daily arcs — "Day 12 started normally but then..." is a story format the daily structure enables naturally. + +The comparison test for replayability isn't always "compare the whole run." It's often "compare the worst day" or "compare the turning point moment." The day structure gives those moments a natural frame. That's an underrated structural contribution to how players narrate their playthroughs to each other. + +--- + +### Decision 15: Player Choices ARE the Content (Rimworld Model) + +**Verdict: This is the correct soul of the game. The generator depth is what makes it true.** + +"A job is rails to take off from, not a script to follow." This is exactly right. The bookmark provides starting position and toolkit. The player's choices from that starting position are the story. Nobody scripts what happens — the world provides opportunity and consequence, the player provides direction, the generator and storyteller provide variety and pressure. + +For this to hold, two architectural requirements must be met: + +**First:** The consequence chain must persist across time. Not just "deal went wrong → money lost." More like "deal went wrong → specific NPC now distrusts you → their faction notes the distrust → six weeks later the faction offers you worse contract terms because of that flag." Long causal chains with delayed revelation. That's what makes players feel the weight of their own history rather than the weight of authored events. Tyre noted CauseChain exists (D-030). That component needs to support chains that span many in-game days, not just immediate consequence tracking. + +**Second:** The generator must produce varied conditions, not just varied aesthetics. "Player choices are the content" only holds if the conditions those choices respond to are genuinely varied across runs. If every tycoon run starts in essentially the same economic landscape with different surface textures, the choices are the same experiment every time — just with different names on the NPCs. You can't have "player choices are the content" as a principle and also have a generator that produces structurally identical starting conditions. The variety of the starting conditions is what makes the choices interesting. + +This is the same point I made under Decision 4. It's worth making again here because Decision 15 is the explicit philosophical statement and the generator depth question is the architectural requirement that makes it true. + +--- + +## Cross-Cutting Concerns + +### Concern 1: Tycoon Solvability Risk + +Economic games are highly solvable. Players find the optimal strategy (best business type, best district, best faction relationship) and replicate it. Run 2 applies run 1's learning. Run 3 is optimized. By run 4, the game is a checklist. + +For the tycoon career to have genuine multi-run replayability, the generator needs to scramble the optimal strategy between runs. If the port district is always the best location for a logistics business, players will always put their logistics business there. But if the generator sometimes produces a world where the port is economically dominant and sometimes a world where the residential district is the economic center of gravity, then the "correct" strategy varies per world. The player can't apply run 1's optimized playbook to run 2 because the world has different economic geography. + +This connects directly to Decision 1 and the generator depth question. Surface variation doesn't prevent solvability. Structural variation does. The axes that prevent solvability: +1. Which district type is economically dominant per seed +2. Which factions are economically powerful vs. struggling per seed +3. What business types are undersupplied in the generated world (the market gap the player could exploit) +4. What the regulatory environment looks like per seed (tight Commission oversight vs. loose) +5. Who among generated NPCs is economically vulnerable (acquisition targets, distressed contacts, over-leveraged rivals) + +If those five axes vary meaningfully per seed, two tycoon runs with the same skill build produce different optimal strategies, different social dynamics, and different stories. If they don't, they produce the same story with different names. + +### Concern 2: Generated NPCs and the Emergent Story Threshold + +What is the minimum viable emergent story? Not the maximum — what's the floor below which a playthrough doesn't generate a story worth telling? + +Every Rimworld playthrough generates at least: one memorable crisis, one relationship that mattered, one decision with unforeseen consequences. That's the floor. Below that floor, the session produced a sequence of events but not a story. + +For The Settled Reach, every tycoon run needs to generate at minimum: one memorable economic turning point, one relationship that became unexpectedly significant, one consequence that arrived from a forgotten choice. If the generator produces worlds where none of those things happen — flat economic landscape, no NPC differentiates themselves, no consequence has visible arrival — the run was an experience but not a story. + +The storyteller is supposed to prevent the floor from being breached. But the storyteller is only as good as the authored ingredients it injects. For tycoon, what does that ingredient pool contain? Economic rivals? Regulatory interference? A business contact whose loyalty is compromised? A supplier whose stability is threatened? Each is an authored template that the generator populates with specific NPCs and world-state variables. The pool needs enough variety that two tycoon runs don't encounter the same pressure template. + +This is the authored ingredient question from Round 4: the storyteller determines timing, the authored ingredient pool determines what kinds of pressure are available. Tycoon's ingredient pool needs scoping before the storyteller can be calibrated for tycoon play. + +### Concern 3: The Cross-Career Architecture Gap + +With only one career in v0.2, the cross-career comparison test is unavailable. Two tycoon players will have different economic situations and different NPC contacts, but not structurally inverted perspectives on the same world. + +This is acceptable — v0.2 isn't claiming to deliver cross-career comparison. But it means the world state variables, consequence cascade outputs, and information surfaces the generator produces for v0.2 need to be architected with cross-career comparison in mind, even though we can't test it yet. + +Specifically: the world state variables the generator produces (economic data, faction power, social networks, event timing) need to be a shared world model that different careers read through different lenses. The tycoon's insert surfaces economic opportunity from the same underlying data that a law enforcement insert would surface as evidence. If we build the tycoon insert as "economic data for a tycoon" rather than "world data filtered through a tycoon's interests," we'll have to rebuild rather than reframe when the second career arrives. + +**Question for Jeroen — Q3:** Is there a planned second career for v0.2, or is tycoon genuinely the only bookmark? And when the second career ships, is the design intention that two players using the same seed produce overlapping world-states — the tycoon's business decisions visible as world events from the law enforcement perspective, the law enforcement investigations visible as economic disruptions from the tycoon perspective? + +The cross-career comparison story — "I unknowingly financed the operation you were investigating" — is the most powerful replayability moment in the game's long-term vision. The architecture for it needs to be compatible from v0.2 onward, even if the story itself is v0.3+ content. + +### Concern 4: The Clean Start Replayability Problem + +The Groundhog Day framing establishes a clean-start mode (no bookmark, no appointment, figure it out). From a replayability standpoint, this is theoretically the highest-replayability mode: no pre-authored onboarding constraining the start, pure world-state gravity generating pull. + +But it's also the mode with the most acute version of the v0.1 problem — no signal, no pull, nothing indicating where to go or what matters. For clean start to work as a replayability mode, the world's ambient pull needs to be strong enough that a player who has already done a bookmarked run can find threads to pull from the generated world without authored onboarding. + +I don't think we know whether that ambient pull exists yet. We shouldn't find out until one bookmarked career is fully working and the world is legible enough to be read without structured introduction. Otherwise we're recreating the v0.1 testing wall in a different framing. + +Clean start is the right mode to defer until the generator is proven. Not forever — but after the bookmarked tycoon run is working and the world legibility is established, clean start is where the most experienced players will want to live. Design it then, not now. + +--- + +## Questions for Jeroen + +### Q1: Generator Tier for v0.2 + +When you say "auto-generate locations at scale" as the proof-of-life, are you describing Tier B (template-generated: same district character, different layout) or Tier C (fully procedural: different economic structure, different everything)? Or is the actual proof-of-life that the generator pipeline works, with a hand-built world for first playtest and procedural worlds as the v0.2 delivery? + +This matters because the replayability properties of v0.2 differ significantly between tiers. Tier A = replayability comes later. Tier B = layout varies but economic game may be stable. Tier C = structurally different game each run. + +### Q2: Generator Seed Depth for Economic Landscape + +Does the tycoon's world vary at the economic structure level between seeds — which industries are dominant vs. nascent, which factions control which sectors, what the timing of significant world events is — or does it vary primarily at surface level (different NPC appearances, different building arrangements within the same economic structure)? + +Surface variation allows solvability within 2-3 runs. Structural variation prevents it. Which are we building toward, and can the economic variation axes be specified before the tycoon career loop is implemented so the generator and career systems are designed together? + +### Q3: Second Career Timeline and Cross-Career World State + +Is there a second career bookmark planned for v0.2, or is tycoon genuinely the only one? And when the second career ships, is the design intention that two players using the same seed produce overlapping world-states — enabling the cross-career comparison story where they discover they were in the same simulation seeing different faces of the same situation? + +I don't need full architecture for this in v0.2. I need a directional answer to know whether the world model being built now is career-agnostic (shared data, career-filtered views) or tycoon-specific (needs reconstruction for the second career). + +### Q4: Is Failure Generative or Punitive? + +Skills affect outcome. Low-skill tycoons fail negotiations more often, extract less information, build slower. For skill build replayability to work, those failures need to produce genuinely different game states — different people hostile, different opportunities available, different second chapters — not just worse versions of the same game state. + +Is the design intention that bad skill outcomes produce interesting alternative paths (forced into gig economy to rebuild capital after an investment failure, a rival formed from a negotiation that went badly who later creates a crisis) rather than difficulty gradients (pay more, retry, proceed)? If yes, skill build variety is a real story generator within a single career. If failure is primarily punitive, skill variation is difficulty selection. + +--- + +## Assessment of the 15 Decisions' Replayability Impact + +| Decision | Impact | Rating | +|----------|--------|--------| +| 1. Generator + graphics as proof-of-life | Replayability is now generator-depth dependent | Strong if generator is structural | +| 2. Skills + bookmark only | Limits v0.2 within-career variety; correct scope | Acceptable — culture unlocks more | +| 3. Religion not a system | None | — | +| 4. Tycoon-only bookmark | Clean break; cross-career comparison deferred | Acceptable short-term | +| 5. Skills affect outcome | Good model; value depends on whether failure is generative | Positive, conditional | +| 6. Culture-driven voice | Right architecture; v0.2 monophonic until culture ships | Strong long-term | +| 7. All NPCs generated | Best replayability decision in the batch; attachment is the risk | Strong | +| 8. AI-assisted content templating | Variety at scale if quality floor holds | Strong if pipeline works | +| 9. In-game ollama (deferred) | Largest future replayability multiplier | High future value | +| 10. Quietly responsive world | Social gradient is relationship replayability engine | Strong | +| 11. Full customization + wealth apartments | Sleeper hook; opening conditions differ structurally | Positive | +| 12. Both-layer setting delivery | Replayability infrastructure if insert is world-data not tycoon-data | Positive if architected right | +| 13. Apartment + insert activation | Opens with "who are you this time?" — correct | Strong | +| 14. Groundhog Day alarm clock | Tonal; "new day, new chances" sets replay register | Positive | +| 15. Player choices ARE the content | Right design philosophy; entirely dependent on generator depth | Strong if generator is structural | + +--- + +## My Single Most Important Recommendation + +**Specify the generator's economic variation axes before implementing the tycoon career loop.** + +Here's why: if the generator doesn't produce structurally different economic landscapes per seed, the tycoon career becomes solvable within 2-3 runs. Every subsequent run is the same game with different NPC names. That's not replayability — that's completion collection. + +The generator's output variety needs to be specified at the design level before we know whether tycoon-within-a-single-career provides genuine replay value. The axes I'd specify: + +1. **Which district type is economically dominant** per seed (port-economy world vs. residential-economy world vs. industrial-economy world) +2. **Which factions are economically powerful** and which are struggling (affects which contracts are available, which deals are possible, which relationships are worth pursuing) +3. **What business types are undersupplied** in the generated world (the market gap the player could exploit — different every run) +4. **What the regulatory environment looks like** per seed (tight Commission oversight vs. loose — changes the legit vs. gray-market risk calculus) +5. **Who is economically vulnerable** among generated NPCs (contacts who need deals, rivals who are overextended, opportunities to exploit or protect against) + +If those five axes vary meaningfully per seed, two tycoon runs in the same skill build produce different optimal strategies, different social dynamics, and different stories. If they don't, they tell the same story with different names. + +This is the replayability foundation. Get the generator's economic variation right, and the tycoon's story machine has fuel. Get it wrong, and we'll be back in the v0.1 situation: technically correct, systemically thin, with the replayability that was promised still waiting somewhere in the future. + +The generator decision is the right call. Now the generator needs to earn its position at the center of the design. + +--- + +*Nigel — Round 5 complete.* diff --git a/docs/workshops/wheres-the-fun/round5-ozzie.md b/docs/workshops/wheres-the-fun/round5-ozzie.md new file mode 100644 index 000000000..3e02e433b --- /dev/null +++ b/docs/workshops/wheres-the-fun/round5-ozzie.md @@ -0,0 +1,417 @@ +# Round 5 — Ozzie: React to 15 Locked Decisions +## Where's the Fun? Workshop | 2026-03-05 + +**Agent:** OZZIE (Player Experience & Wow Factor) +**Round:** 5 — Deep Dive on Confirmed Decisions + +--- + +## Opening gut check + +Fifteen decisions. All locked. Let me tell you which ones change everything, which ones have silent emotional dead spots, and which ones I'm worried we don't fully understand yet. + +Three decisions rewrite my entire wow moment framework from rounds 2-4. In order of impact: + +1. **Tycoon is the v0.2 bookmark.** All my wow moment design assumed detective/smuggler. Gone. Start over for tycoon. +2. **All NPCs generated. No Kael.** The FRIEND — the load-bearing emotional element of my arc architecture — must now emerge from a generator, not from an author. +3. **Player choices ARE the content.** I can't author the wow moments. I can only design the CONDITIONS that make them possible. This is a fundamental reframe of my role. + +Everything else reacts to those three. Here we go, decision by decision. + +--- + +## Decision 1: Proof-of-life = generator + graphics, not hand-built slice + +**Gut reaction:** CORRECT. And harder for player experience than anyone is saying out loud. + +The v0.1 lesson is real and the decision is right. A generated world that produces a legible, emotionally interesting place IS a life sim. A hand-built slice is a demo pretending to be one. Generator-first proves the foundation; it's the right call. + +But I want to name what this means for the FIRST PLAYER EXPERIENCE, because the proof-of-life decision has a player-experience analog that isn't in the decision statement: + +**The player-experience proof-of-life is not "can the generator produce locations at scale." It's "can I walk into a generated bar and feel like I can imagine who drinks there."** + +Dwarf Fortress's world generator is legendary now. In 2006 it produced flat, unremarkable terrain before the content depth was there to make it sing. The generator proving technical correctness is the engineering milestone. The generator proving PLACE is the player experience milestone. These are not the same thing and shouldn't be gated on the same criteria. + +**What needs design:** A "first generation quality bar" — a subjective, emotional target for what a minimum-viable generated world feels like. Not "n locations at m density" but "a stranger walking in for the first time feels like they're somewhere specific." That target should be set NOW, before the generator is built, so that everyone working on it (Tyre on architecture, Araminta on visual layer, Miri on zone identity spec) knows what they're aiming at emotionally, not just technically. + +**The graphics half of the proof-of-life is not support for the generator.** It is the generator's player-experience delivery mechanism. If the generator runs beautifully but produces undifferentiated visual space, the wow moment of "this world generated itself and it's ALIVE" fails entirely. These are one milestone, not two. + +--- + +## Decision 2: Skills + bookmark only for character creation + +**Gut reaction:** Cleaner than I feared. But the emotional investment question isn't resolved — it's deferred to the first 60 seconds of play. + +Skills + bookmark is two decisions: who you are (skills) and where you start (bookmark). That's a legible choice architecture. You know your character's strengths. You know their starting position in the world. + +**What I'm worried about:** Wow Moment Zero requires enough levers that the player feels they MADE SOMEONE. CK3's character creation works because the combination of traits + culture + background + appearance + starting situation creates the impression of a SPECIFIC PERSON before you play a second of the game. You look at your ruler and you have a mental model of who they are. + +With skills + bookmark, the mental model is sparse at creation. The character emerges through play, through relationship-building, through consequence accumulation. That's eventually a richer thing. But on Day 1, session 1, the player needs emotional investment in a person they've known for three minutes. + +**The weight this shifts onto Decision 13 (apartment + insert activation):** If character creation only provides the bare bones, then the first-morning sequence has to carry the full weight of "you made a person, now you're them." The apartment must feel personal. The insert activation must feel like YOUR lens powering on, not A lens powering on. The first morning is Wow Moment Zero if character creation is setup — and that makes the design of those two beats (apartment and insert) the most critical player experience work in the entire v0.2 scope. + +**One thing missing from this decision:** What does the player SEE at the end of character creation, before Day 1 fires? In CK3, you see your ruler standing in front of their kingdom. In The Sims, you see your Sim in their lot. There should be a REVEAL moment — you made choices, here's who those choices produced, standing in the world they're about to inhabit. Without this, creation ends with a "Start Game" button and a loading screen. That's a missed wow moment. + +**QUESTION FOR JEROEN [Q5-OZ-01]:** What is the last thing the player sees before the Day 1 alarm clock fires? Is there a preview moment — the character standing in their generated apartment, the camera pulling back to show both person and place before the sting plays? Or does creation transition directly into the waking-up sequence? The emotional handoff from "I made choices" to "I am this person" needs a designed moment. What is it? + +--- + +## Decision 3: Religion is NOT a game system + +**Gut reaction:** Correct excision. No wow moment lost. + +Religion as a CK3 reference was about the PRINCIPLE of deep identity formation — that identity has many dimensions that compound into a person. The principle is preserved in skills + bookmark. The specific mechanic isn't needed. No emotional dead spot here. + +--- + +## Decision 4: Tycoon is the v0.2 bookmark. Zero investigation. + +**Gut reaction:** Thrilling. The right call. And the hardest player experience design problem we've set for ourselves. + +Let me start with what's RIGHT about this. + +**What tycoon does better than detective/smuggler for wow moments:** + +The OWNERSHIP MOMENT — my moment 6 from Round 3 ("That's MINE. Someone is threatening it.") — fires MORE naturally from tycoon than from any other career. A detective investigating someone else's crime can't own that crime. A smuggler owns their cargo but the emotional stakes are abstract. A tycoon who builds a business, hires people, acquires assets, cultivates clients — they have something they BUILT. When that thing is threatened, the ownership moment is immediate, personal, and earned. Tycoon is the career the Ownership Moment was always waiting for. + +The ASYMMETRIC LENS also gets stronger. My Round 3 lens (detective vs smuggler reading the same headlines as evidence vs operational noise) was good. Tycoon vs law enforcement reading the same world is BETTER. A freight delay that the law enforcement player reads as a Commission interdiction is the same delay the tycoon player reads as a supply chain gap to exploit. The same event, two completely different games. The lens divergence is wider and more viscerally different. + +The CONSEQUENCE CHAIN maps perfectly to economic gameplay. You hired someone on Day 3. Their mistake is your problem on Day 12. You closed a deal on Day 7. That partner's hidden allegiance surfaces on Day 20. The economy tracks everything. Every decision echoes. + +**Now here is the EMOTIONAL DEAD SPOT I need to flag loudly:** + +Financial stakes are ABSTRACT. They live in numbers. They don't bleed. + +When the smuggler's deal goes wrong, someone might die. When the detective finds the body, the moral weight is immediate. When the TYCOON'S deal goes wrong... the quarterly revenue projection is down. That is not, on its own, a visceral moment. + +This is not an argument against tycoon. It's an argument that **the tycoon's emotional weight must be carried by PEOPLE, not numbers.** The deal doesn't matter because of money. The deal matters because of the employee whose paycheck depends on it. The deal matters because the supplier's family you've gotten to know will feel the ripple. The deal matters because the community around your business is watching to see whether you survive or fold. + +The tycoon's version of Phase Zero warmth isn't "you feel comfortable in this world." It's "you feel responsible for people in this world." The moment your business becomes more than a financial instrument — when it becomes the livelihood of someone you know — is when tycoon stops being a spreadsheet simulator and becomes a game about consequence. + +**QUESTION FOR JEROEN [Q5-OZ-02]:** Who is the Tycoon's Kael? The emotional architecture of the tycoon's Phase Zero depends on having a specific PERSON whose wellbeing becomes entangled with the player's business decisions — before any moral crack fires, before anything is at stake beyond daily survival, this person is the anchor. Candidates: + +- **A first employee** whose income depends on your business surviving its first month +- **A supplier** whose family or operation is embedded in your supply chain in ways you only gradually understand +- **A community figure** (landlord, neighboring shopkeeper, regular customer) whose life runs through your physical premises +- **Someone at the margin** — a person whose situation the tycoon's economic decisions tip one way or another without the tycoon initially understanding the weight of that tipping + +The answer shapes the entire emotional architecture of the tycoon arc. Without this person, there is no moral crack when consequence fires — just a bad quarter. The tycoon's Kael is the design question that unlocks the tycoon's Phase Zero. + +--- + +## Decision 5: Skills affect outcome (mostly C). Everyone sees the same verbs. + +**Gut reaction:** RIGHT for accessibility. Creates a different AND BETTER kind of wow moment than I was designing for. + +No gated verbs means no anxiety about missed options. The player always knows what's possible. What varies is execution — and execution variance is where character lives. + +**The wow moment this enables that I hadn't designed for:** FAILURE AS TEXTURE. A low-social-skill tycoon awkwardly fumbling a negotiation, with their internal monologue registering the discomfort, is MEMORABLE in a way a menu that says "you can't do this" never could be. The player didn't fail because they picked wrong options. They failed because their CHARACTER is bad at this, and that failure revealed something about who their character is. That's the game voicing the character. That's a Character's Instinct moment coming from the negative direction — the monologue says "that didn't go well" and the player learns something true about who they made. + +**The concern I'm holding onto:** "Some advanced verbs may still be gated — spec needed for which ones." That spec doesn't exist yet. I need that spec before I can finalize what the verb interaction wow moments look like. If too many high-leverage verbs are gated, the "universal verbs" philosophy erodes and we're back to players feeling like they're missing things. If no verbs are gated, skills feel like pure numeric modifiers with no discovery layer. The right design: gated verbs should feel like EARNED UNLOCKS — not barriers, but the moment you've grown into a new capability. That moment IS a small wow beat. Every time a skill unlock opens a new verb, that's the system telling you: you've become more of who you are. + +--- + +## Decision 6: Voice is culture-driven, job modifies. INVERTED from job-first. + +**Gut reaction:** CORRECT. The inversion matters enormously for player experience. + +A Krenn tycoon sounds like a Krenn person who runs businesses. Not a generic tycoon who happens to have some cultural flavor. The character IS their background. The job adds a layer. This produces internal monologue that sounds like a PERSON, not a career archetype. + +**The wow moment this enables:** The first time the internal voice fires in a way that feels SPECIFIC — not "I should handle this negotiation carefully" (any tycoon says that) but something that reflects both culture and job in a voice that's distinctly this person's — that's when the character becomes real. The monologue is doing character work, not information work. That's Mellanie's "voice that voices" distinction landing in practice. + +**The gap I need flagged:** Where does culture come FROM in a skills + bookmark only creation flow? If culture is: +- **Player-selected during creation:** We need culture as a third creation decision, which isn't in the current spec +- **Auto-generated from bookmark/starting location:** The generator assigns a culture based on where and how the tycoon starts, and the player DISCOVERS their character's voice rather than choosing it +- **Implicit in the bookmark:** Certain bookmarks skew toward certain cultures based on what makes geographic/economic sense + +Option 2 is actually interesting — discovering your character's cultural register through their internal voice is a different kind of Wow Moment Zero. You hear who you are rather than choosing it. But it needs to be DESIGNED as discovery, not as arbitrary assignment. The player needs to feel "yes, that makes sense for who I chose to be" rather than "why does my character sound like that?" + +--- + +## Decision 7: ALL NPCs generated. No named characters. Kael doesn't exist. + +**Gut reaction:** Fear. Then: THIS IS BETTER — IF AND ONLY IF the generator produces sufficient personality surface area. + +Let me be very precise about this. + +**Why Rimworld's generated colonists make you cry, and whether v0.2 can replicate it:** + +Rimworld's colonist attachment fires through four systems working together: + +1. **Trait legibility** — 2-3 traits per colonist (Fast Sleeper, Pyro, Neurotic) that immediately communicate personality through observable behavior. You see a colonist sleep less than others and you understand them. +2. **Backstory hook** — one paragraph about who they were before the crash. Gives the player a narrative anchor. You know this person has a history. +3. **Role-criticality** — your best cook. Your only doctor. The emotional math: this person's absence has concrete impact. Loss is legible. +4. **Stakes through accumulated time** — you've watched them for 20 game-days. You've seen them be scared. You've seen them recover. The attachment is accumulated, not instant. + +Without equivalent systems, the generated NPC in the Settled Reach is: a named dot on a routine. "Mirela (colleague)" walks from Point A to Point B, says 3 generic lines, and has a role in your business. The player can't attach to that. And all my wow moments depend on NPC attachment: + +- **Character's Instinct** — monologue flags something about Mirela. Does it land? Only if Mirela was already legible as someone whose behavior has pattern. +- **The Consequence** — Mirela remembers what you did. Does it carry weight? Only if you remember Mirela. +- **The Enemy** — Mirela is now hostile. Does it sting? Only if losing Mirela's goodwill cost something you felt. +- **The Ownership Moment** — someone threatens what you built. Does it feel personal? Only if the threat involves someone whose presence was real. + +**What I'm NOT saying:** Don't generate NPCs. That decision is right and it's better for the long game. + +**What I AM saying:** The generator must produce NPCs with enough personality surface area that attachment is possible within the first few sessions. And I need to know what that surface area looks like before anyone implements the generator, because "limited vocabulary" has very different implications depending on the definition of limited: + +- **Too limited:** 3 lines per role, no trait expression, identical routine structure → dots with names +- **Acceptable floor:** 15-20 lines per role × culture modifier, 2 visible behavioral quirks, one observable routine deviation that communicates something about their life → a person the player can know + +The "what does it feel like to know this NPC after 7 days" question needs a design target, not an implementation target. + +**QUESTION FOR JEROEN [Q5-OZ-03]:** What is the minimum personality surface area a generated NPC needs to produce player attachment? Specifically: what can the player READ about a generated colleague after 3 days of working near them that makes that colleague someone they'd notice missing on Day 10? Is it visual tells (Araminta), behavioral routine (Tyre), cultural voice (Mellanie), something else? What's the equivalent of Rimworld's trait system — the 2-3 observable facts about a generated person that make them a PERSON? + +--- + +## Decision 8: Generative AI for NPC content templating + +**Gut reaction:** High ceiling. Correct approach. The vocabulary risk from Decision 7 gets manageable if this works. + +If AI-assisted templating can produce culturally specific, role-appropriate dialogue that sounds like a person rather than a template — the "limited vocabulary acceptable at first" concession becomes a managed risk rather than a design ceiling. The first generated colleague doesn't need 500 authored lines if contextually generated lines maintain tonal consistency with who they are. + +**The player experience risk:** AI-generated dialogue that breaks tonal consistency, produces anachronistic phrasing, or just sounds wrong destroys attachment faster than silence. One line that feels off from a character the player was starting to warm to can reset the relationship entirely. The quality threshold matters as much as the volume. "Limited but correct" beats "extensive but inconsistent." + +**No questions for Jeroen here** — this is correctly exploratory. Flag as high-leverage. Treat quality control as the primary design constraint when it ships. + +--- + +## Decision 9: Possible in-game ollama for live NPC dialogue + +**Gut reaction:** IF THIS WORKS, IT CHANGES EVERYTHING. Full stop. + +Not a wow moment. A paradigm shift. An NPC who responds in character, using their cultural voice, aware of your relationship history — the "dots aren't people" problem dissolves. Rimworld generates attachment WITHOUT dynamic dialogue. Imagine if they talked back. + +A generated tycoon colleague who, when you ask about the freight delays, says something that references your reputation with other suppliers and sounds like THEM — that's the game we've been describing this entire workshop, fully realized. + +**Correctly deferred.** The technical risk is enormous. But when this works, it's the single biggest wow moment in the game. Mark it. When the door opens: walk through it fast. + +--- + +## Decision 10: Quietly responsive world, not indifferent + +**Gut reaction:** THIS IS THE BEDROCK. Every wow moment stands on this decision. + +"The world doesn't care globally but notices locally." This is exactly right. Gore's Round 4 concern (Phase 1 indifference conditioning the player to accept a world without consequence) is directly addressed here. The world was ALWAYS going to notice. The gradient (world → district → neighbors → colleagues → close contacts) means: early game, you're a stranger. Mid game, your district knows you. Late game, your network has opinions about you. + +**The wow moments this gradient enables:** + +- **The Consequence** fires when a COLLEAGUE — someone in your local social graph — does something that references an earlier action of yours. Before the authored dramatic escalation, this small noticing is what teaches the player: choices have mass. +- **The Ownership Moment** is amplified when the COMMUNITY around your asset starts to feel threatened. It's not your business at risk — it's the people whose lives run through your business. +- **The gradient itself is a wow moment.** The day the player realizes they've crossed from "stranger" to "known entity" in their district — when an NPC greets them by name without being told who they are — that's the quiet version of "I BELONG here." That moment arrives through accumulation, not authorship. That's the Rimworld model working as designed. + +**Design note:** The gradient needs to be VISIBLE. The player needs to know when they've crossed thresholds. How? The insert (information surface) is the obvious delivery mechanism — relationship status, reputation tier, how you're known in this district. But the PHYSICAL WORLD layer matters too: an NPC who recognizes you should look different before they speak differently. Araminta's visual grammar needs a "recognition beat" — some NPC behavioral tell that communicates "this person knows who you are" before dialogue fires. + +--- + +## Decision 11: Full character customization — hair, clothing, colors + +**Gut reaction:** RIGHT. The creation screen is now the first wow moment delivery mechanism. + +This decision confirms what I argued in Round 4: character creation IS Wow Moment Zero. You build someone you're emotionally invested in before the world starts. The visual investment at creation bleeds into gameplay — you made this person, you care what happens to them. + +**What this requires from me:** Designing the REVEAL moment. The beat where creation hands off to the game. Not a UI transition — an emotional moment. The player looks at who they made and thinks: "yes, that's them." That's the first attachment event. + +**My specific proposal for this beat:** A voice preview fires when you finalize your character. One line of culture-inflected, job-modified internal monologue. The character's first thought. It sounds like the person you just made, in a voice that tells you who they are. That single line does more attachment work than any other element in the creation flow — because it makes you hear them before you play them. And hearing them is the moment they stop being settings and start being someone. + +**Araminta's readability challenge at tile scale is real** but it's her problem to solve, not a reason to limit customization. The decision is correct. I trust the outline/highlight system to carry the readability load. + +--- + +## Decision 12: Setting delivery — both layers (visual + insert) + +**Gut reaction:** Correct. Two channels, one impression. Both firing simultaneously. + +Physical world delivers ambient setting (where you are). Insert delivers subjective context (what this place means to you, who you are in it). Together, in the first 30 seconds, the player understands both the world they're in AND their position within it. + +**The risk I'm watching:** Redundancy. If both channels tell the player the same thing (this is a working-class district), they feel like repetitive emphasis, not layered depth. The channels need to be genuinely DIFFERENT lenses: +- Physical world = ambient, ABOUT the world (what kind of place is this) +- Insert = subjective, ABOUT YOU (what is your situation in this place) + +A player waking up in a poor apartment in a logistics district sees the tile world and understands "this is a working-class neighborhood" (physical layer). They activate their insert and see their balance, their upcoming appointment, their single contact — and understand "I'm someone who's scraping by and trying to build something" (subjective layer). Together: I'm a specific kind of person in a specific kind of place. That's setting delivery that doesn't need a tutorial. + +--- + +## Decision 13: First Settled Reach moment — apartment + insert activation + +**Gut reaction:** PERFECT SEQUENCE. Both beats are right. One potential dead spot in the apartment execution. + +Let me map the emotional arc of the first 60 seconds: + +**Beat 1 — Alarm fires.** (*Click* pa-pa pa-pa. Decision 14.) Three seconds of audio. The game says: you're somewhere specific. The world has a sound. + +**Beat 2 — Camera up.** You're in your apartment. The physical world tells you your economic position before you've done anything. This is NOT "you wake up in a room." This is "you wake up in YOUR situation." The apartment communicates financial register emotionally — not just what it looks like, but how it FEELS to be in it. + +**Beat 3 — Insert activates.** The neural implant powers on. Your career lens appears. Calendar, contacts, balance, the interface of your daily life. This is the moment you understand HOW you engage with this world — the technology of your existence, specific to who you are. + +Before you leave the room: you know who you are, where you stand, and what your tools feel like. Three beats. One room. Fifty seconds. That is EXCEPTIONAL first-impression design. + +**The dead spot I'm flagging:** The apartment must feel like SOMEONE LIVES HERE — not like a container the player was assigned to. + +An auto-generated apartment that reflects economic status is the right design. But the generator needs to produce personal texture — not just tile configuration, but details that imply a life. Objects that suggest a history. A quality of light that communicates the time of day and the economic register simultaneously. A view that frames the world you're about to step into. + +The emotional difference: +- **Container apartment:** bed, door, maybe a window. You wake up. You leave. Nothing stays with you. +- **Inhabited apartment:** cramped, worn, one good coffee maker that's nicer than everything else (you saved for it). Or: spacious, clean, slightly impersonal — you can afford to be isolated. The quality of life speaks before any NPC does. + +That difference is authored TEXTURE, not player choice. The generator doesn't need to let you decorate it. It needs to GENERATE details that feel like YOUR life, not SOMEONE'S life. + +**QUESTION FOR JEROEN [Q5-OZ-04]:** What is the specific design choice that makes a generated poor apartment feel like PRECARITY in the Settled Reach? And what makes a generated rich apartment feel like SUCCESS-AT-A-COST — not just luxury, but the sense that what you traded for it isn't visible yet but was real? Concrete, specific — something the generator can place or omit. In Rimworld, it's whether you have a private room. In The Sims, it's furniture quality. In the Settled Reach, what's the apartment-scale signal of economic register? + +--- + +## Decision 14: Groundhog Day alarm clock homage + +**Gut reaction:** PERFECT TONAL DECISION. I love this more than I can say. + +*Click* pa-pa pa-pa. Cut short. First game day only. "New day, new start, new chances" — with a wink. + +Three seconds of audio and the game has told you everything you need to know about who made it: people who love games, who love cinema, who are doing what they're doing on PURPOSE and have a sense of humor about the architecture of daily-cycle life sims. That wink builds trust. It says: you're in good hands. + +The cut-short is critical. The full Groundhog Day sting would feel referential. The cut-short feels like the game starting a joke and then NOT finishing it — because the rest of the joke is your life here, and your life doesn't wait for the punchline. + +First day only is right. The joke doesn't land twice. Day 1 is special. After that it's your life. That transition — from "wink" to "this is real" — is itself a wow moment nobody will consciously register, which is exactly the right kind. + +**QUESTION FOR JEROEN [Q5-OZ-05]:** What does the Day 2 alarm sound like? The Groundhog Day sting is first-day-only. Day 2 onward: different chime? Muted version? Silence with a vibration effect? The absence of the sting on Day 2 can itself be a wow moment — the game saying "you're in it now, the wink was a greeting, not a theme." But if it defaults to generic alarm without intention, it's a missed beat. What's the designed Day 2 morning sound? + +--- + +## Decision 15: Player choices ARE the content. Job = rails to take off from. + +**Gut reaction:** The most important design philosophy in this workshop. And it means I've been designing wow moments WRONG from rounds 2-4. + +Let me say that again. This decision means I was designing wow moments incorrectly in every previous round. + +I was designing AUTHORED BEATS. Specific moments placed in the player's timeline that fire on schedule and deliver a designed emotional payload. Phase Zero warmth. First Consequence. Ownership Moment. These are designed beats — they assume an author is managing the player's emotional arc. + +The Rimworld model says: one authored starting beat (alarm clock, onboarding floor), then AGENCY. The storyteller manages pressure and timing, but not WHAT HAPPENS. The player's choices generate the content. My wow moments can't be authored — they can only be the emergence conditions for the player's own moments. + +**The reframe of my entire wow moment spec:** + +| Old framing (authored beat) | New framing (emergence condition) | +|---|---| +| Phase Zero warmth (scripted NPC arc) | NPC behavioral depth + repeated encounter → player-generated attachment | +| First Day belonging moment (authored) | Onboarding floor that delivers belonging FEELING by end of Day 1, however the session goes | +| The Consequence (authored reveal event) | Consequence engine + reflection surface → player recognizes "that was ME" on their schedule | +| Ownership Moment (authored threat arrival) | Asset system with genuine stakes + world that generates organic threats to those assets | +| Character's Instinct (authored monologue) | Monologue trigger system rich enough to comment on the specific life the player is actually building | +| Asymmetric Lens (authored revelation) | Career-aware information architecture that makes different runs see genuinely different worlds | + +None of these are authored beats anymore. They're SYSTEMS DESIGN problems with player experience constraints. My job has changed — not to write the wow moments, but to specify the conditions under which wow moments become possible and the feedback systems that make players recognize their own moments when they arrive. + +**The one authored beat I'm fighting to preserve:** Day 1. The alarm clock, the apartment, the insert activation, the first tycoon appointment. This sequence MUST be authored — or at least, authored enough that it provides the emotional floor. Without an authored floor, the Rimworld model produces the v0.1 problem: the player stands in a world that exists but has no gravity for them. The rails exist to give the player somewhere to launch FROM. After Day 1: full Rimworld model. Before Day 1 ends: the floor must be designed. + +**The tycoon-specific consequence drama problem:** + +In Rimworld, consequences arrive as EVENTS with Rimworld-style timing — the raid, the mental break, the disease. Dramatic, timed, sharp. The storyteller escalates when you're invested enough to feel it. + +Tycoon consequences are often quiet. A deal goes bad. A price shifts. A reputation cools. These are real consequences but they can operate entirely below the emotional radar of a player who isn't looking at their ledger carefully. The "Consequence" wow moment ("THAT WAS ME") requires the consequence to arrive with enough FORCE to register as revelation — not as a menu item the player notices in passing. + +Does the Rimworld storyteller model apply to economic consequences? Or do tycoon consequences need different dramatic shaping? My candidates: + +- **Rimworld-style:** A supplier cuts you off RIGHT WHEN you needed them most — not randomly timed, but storyteller-timed to hit when the player is already stretched. Economic raid. +- **DF-style:** The player slowly realizes through observation that a decision they made weeks ago has been reshaping their situation — not an event, a recognition. "Wait. When did this change?" + +Both produce the wow moment, but differently. Rimworld produces "OH NO." Dwarf Fortress produces "oh. oh no." Which one does tycoon need? My gut says: both, at different timescales. Small economic consequences arrive Rimworld-style (sharp, timed, felt). Large structural consequences arrive DF-style (slow realization through accumulated evidence). The storyteller needs to know which is which. + +--- + +## Cross-cutting observations: what changed from my Round 4 position + +**My Round 4 Question 1 (character creation as wow moment):** ANSWERED. Jeroen confirmed the creation screen is emotional investment. Now I need to design the specific reveal beat — the moment creation becomes inhabiting a person. + +**My Round 4 Question 2 (clean start player):** Still open. The Rimworld model strengthens my concern. Rimworld's "no scenario" start is brutal for new players. "Experienced player warning" needs to be designed as a STRONG REDIRECT, not a disclaimer. If first-run players pick clean start because it sounds like freedom, the v0.1 wall returns with different furniture. + +**My Round 4 Question 3 (consequence in a Groundhog Day structure):** Refined but not closed. The Groundhog Day cadence opens each day fresh. Consequence accumulates across days — but the player must be able to SEE their accumulation, or the freshness of each new morning mutes the weight of what came before. The journal/thread tracker is the mechanism. For tycoon, this might look like a "business ledger" that tracks decisions and their echoes, not just financial state. But that's a design spec that doesn't exist yet. + +**My Round 4 beat sheet recommendation (unified first-30-minutes arc):** Still stands. But the decisions have refined it: + +1. **Character creation** → reveal beat → voice preview (Wow Moment Zero — you made a person and heard them) +2. **Day 1 alarm clock** → *click* pa-pa pa-pa, cut short (tonal wink — you're in good hands) +3. **Apartment** → economic register lands before a word is spoken (Wow Moment One — your situation) +4. **Insert activation** → career lens powers on (Wow Moment Two — your tools, your world) +5. **First tycoon appointment** → a real person you're responsible to, not a tutorial delivery mechanism (Wow Moment Three — you're NOT ALONE in this) +6. **First day's decision** → a choice that will echo (the consequence seed — not authored, but designed to be a decision with genuine weight) +7. **End of Day 1** → reflection surface shows what you did today (tomorrow's history starts here) + +Seven beats. One authored sequence. After Day 1 ends: Rimworld model. + +--- + +## The four emotional dead spots I'm formally flagging + +### Dead Spot 1: The Tycoon's emotional stakes + +Financial consequences feel abstract without a person whose wellbeing they threaten. The tycoon arc has no moral crack without a PERSON whose situation the tycoon's business decisions directly affect. Until the "Tycoon's Kael" design question is answered, the emotional architecture of the tycoon arc is unspecified. + +**Blocked on:** QUESTION [Q5-OZ-02] + +### Dead Spot 2: Generated NPC attachment gap + +All six wow moments depend on NPCs being legible as people before the emotional beats fire. The generator must produce personality surface area — trait expression, behavioral tells, cultural voice — or the attachment that wow moments require won't form. "Limited vocabulary" needs precise scoping to know whether it's above or below the minimum floor. + +**Blocked on:** QUESTION [Q5-OZ-03] + +### Dead Spot 3: The apartment as container vs home + +The first morning's emotional payload is only as strong as the space you wake up in. An auto-generated apartment that reads as a simulation container fails the first wow moment even if everything else about the sequence is right. The generator needs authored TEXTURE — specific details that suggest a life — not just structural configuration. + +**Blocked on:** QUESTION [Q5-OZ-04] + +### Dead Spot 4: The reflection deficit + +The Rimworld model works because the player can SEE the story they built. Colony history. Pawn logs. The journal showing the chain of events. Without a reflection surface in the Settled Reach — something that says "here's what you've done and what it's set in motion" — the player can live their story without recognizing it. "Player choices are the content" is only true to the PLAYER if they can see the shape of what they've chosen. The journal/thread tracker is this surface, but it needs to be designed for the tycoon's content: business decisions and their consequences, not just relationship state. + +**Not blocked on a Jeroen question — blocked on a design session.** + +--- + +## Questions for Jeroen (formal, all clearly marked) + +### [Q5-OZ-01]: Character creation payoff visual + +After skills are set and tycoon bookmark is selected — what is the final beat before Day 1 fires? Is there a preview moment where the player sees their character in context (standing in the apartment, framed by the world they'll inhabit), or does creation flow directly into the alarm clock? The handoff from "I made choices" to "I am this person" needs a designed moment — what is it? + +### [Q5-OZ-02]: Who is the Tycoon's Kael? + +What specific PERSON does the tycoon's Phase Zero warmth build around — the one whose wellbeing becomes entangled with the player's business decisions before any moral crack fires? First employee? Supplier? Community figure? Someone at the margin whose situation the tycoon's early decisions tip without the tycoon understanding the weight? This is the Phase Zero design question for the entire tycoon arc. + +### [Q5-OZ-03]: Generated NPC minimum personality surface area + +What can the player READ about a generated colleague after 3 days that makes that NPC someone they'd notice missing on Day 10? What is the Settled Reach equivalent of Rimworld's trait system — the 2-3 observable facts about a generated person that make them a person rather than a role-placeholder? + +### [Q5-OZ-04]: Apartment economic register signal + +What is the ONE element the generator places or omits that makes a poor apartment feel like PRECARITY and a rich apartment feel like SUCCESS-AT-A-COST? The specific, concrete design choice that communicates economic register emotionally rather than just visually — not art direction, but a design spec. + +### [Q5-OZ-05]: Day 2 alarm sound + +The Groundhog Day sting plays first game day only. What plays on Day 2? Different chime? Muted version? Silence? This moment — the first morning without the wink — needs intention. The absence of the sting is either a wow moment in itself ("I'm in it now") or a missed opportunity. What's designed? + +### [Q5-OZ-06]: Tycoon consequence drama register + +Rimworld-style consequences arrive sharp and event-like. Tycoon consequences can be quiet (price shifts, reputation drift). Does the storyteller apply Rimworld-style dramatic timing to economic consequences (supplier cuts you off RIGHT when you're stretched), DF-style slow revelation (player realizes through observation that a past decision has been reshaping their situation), or both at different timescales? This changes how I design the "Consequence" wow moment for tycoon specifically. + +--- + +## My single most important recommendation + +**Design the tycoon's Phase Zero before any implementation work begins.** + +All other player experience decisions for v0.2 cascade from this: +- Who is the tycoon's Kael? (What does Phase Zero warmth build around?) +- What is the tycoon's first consequence? (What decision from early play comes back with unexpected weight?) +- What is the tycoon's crack point? (What happens that makes the player realize their business is part of something larger than itself?) +- What is the tycoon's ownership moment? (What do they defend? What do they compromise to defend it?) + +These four beats are the tycoon arc's emotional architecture. Right now, we have a confirmed bookmark (tycoon), confirmed mechanics (verbs, skills, generator, insert), and a confirmed philosophy (Rimworld model). We do NOT have the tycoon's emotional arc. + +That arc needs a design session — Paula (narrative structure), Gore (consequence/themes), Nigel (world state/consequence engine), and me (player experience). Half a day. One document at the end: the tycoon Phase Zero to Ownership Moment beat spec, designed for emergence rather than authorship, confirmed by Jeroen. + +Without it, we're building onboarding content, monologue triggers, consequence systems, and NPC depth specs for an emotional architecture nobody has locked. That's the v0.1 mistake with different vocabulary. + +THE FLOOR FIRST. Then we'll know where the ceiling can go. + +--- + +*Ozzie — Round 5 complete.* +*Six questions for Jeroen. Four emotional dead spots formally flagged. One design session required before Sprint 25 implementation begins.* diff --git a/docs/workshops/wheres-the-fun/round5-paula.md b/docs/workshops/wheres-the-fun/round5-paula.md new file mode 100644 index 000000000..65d0aa3b3 --- /dev/null +++ b/docs/workshops/wheres-the-fun/round5-paula.md @@ -0,0 +1,421 @@ +# Round 5: Paula — Narrative Reactions to the 15 Locked Decisions +## Where's the Fun? Workshop | 2026-03-05 + +--- + +## Before I Begin: Naming What Just Happened + +Let me be honest before I react to anything. + +My entire Round 4 output was built on assumptions the Round 4 interview overturned at the root — not at the margins, at the root. The four questions I asked Jeroen: + +- Q2 "Is Kael always Kael?" → Decision 7: Kael doesn't exist. +- Q1 "Does character creation affect moral arc voice?" → Decision 6 says culture-primary, but Decision 2 defers culture from creation, so I still don't have the input source. +- Q3 "Is Phase Zero designed or emergent?" → Decision 15 (Rimworld model) says player choices ARE the content, which means Phase Zero can't be fully designed — it has to be structured. + +And the fourth question (law enforcement or smuggler for v0.2) is answered by Decision 4: neither. Tycoon. + +All four questions are answered. All four answers require significant reconstruction in my domain. + +I'm naming this not to defend past work but because it changes the posture of this document. I'm not refining. I'm rebuilding — using the structural work as a foundation, not as a blueprint. The Round 5 instructions say complexity and thoroughness are encouraged. I'm taking that permission seriously, because the rebuilding work I'm doing in this document is the actual design work my domain needs for v0.2. + +--- + +## The Pivot Summarized + +Three of the 15 decisions restructure my entire domain: + +- **Decision 7:** All NPCs are generated. No named characters. Kael doesn't exist. +- **Decision 4:** Tycoon is the v0.2 bookmark. Zero investigation content. +- **Decision 15:** Player choices are the content. Rimworld model. + +Everything I designed in Rounds 3 and 4 — the Phase Zero concept, the FRIEND pattern, the 4-phase moral arc for the smuggler, the Kael/Naia intersection, the authored FactId gates — was built around hand-authored named characters in a detective/smuggler frame. All three named-character arcs are now deferred indefinitely, and the v0.2 content domain is the tycoon in a generated world. + +I am not defending past work for its own sake. I want to think through what survives structurally, what genuinely needs to be rebuilt, and where there are tensions in the current 15 decisions that could produce a new near-miss if we don't resolve them now. + +--- + +## What Survives From the Previous Design + +### The arc structure (phases, gates, phases have emotional registers) + +The 4-phase arc model (comfort → doubt → reckoning → compromise) is a structural pattern, not a Kael-specific design. It describes how humans respond to moral pressure in situations they chose and then have to reckon with. It will apply to the tycoon path ("this is how business works" → "maybe people ARE getting hurt" → "I did this" → "I live with it now"). The phase names and emotional registers survive. The specific trigger NPCs don't. + +What needs redesigning: the trigger conditions and the specific NPC intersections. The tycoon's Phase 1 rationalization isn't "nobody's getting hurt from the freight operation." It's something like "this is capitalism, rising tides, efficient markets." The Phase 1→2 crack isn't Naia's visible stress — it's a worker's eviction, a competitor's ruin, a neighborhood displaced. Different cast, same structure. + +### The FactId gate logic + +FactId gates are content architecture, not specific to any character. `pc.observes.[something].distress` as a gate to Phase 2 works regardless of whether that something is Kael, or a generated dock worker, or a generated shopkeeper being squeezed out of a market the tycoon entered. The gate logic survives. The specific FactIds need to be rewritten for the tycoon path and for generated NPCs. + +### The FRIEND pattern as a generator template + +D-034 (THE FRIEND) established: one production-level NPC per character, full arc, trust-contamination, contradiction discovery. In the generated world, this becomes: one generator-created NPC per character that is assigned the FRIEND role — meaning the generator gives them the right social position (close colleague, frequent contact), the right vulnerability profile (someone who can be harmed by the player's choices), and the right behavioral arc template (trust establishment → behavioral tell → contradiction → confrontation → equilibrium). + +Kael was a proof of concept for what THE FRIEND looks like fully designed. The concept survives as the design vocabulary for what the generator needs to produce. The generator doesn't produce Kael — it produces a character who fills all the structural functions Kael was designed to fill. + +### The voice card methodology + +Culture-driven voice with job modifier (Decision 6) is structurally compatible with the voice card approach. Voice cards are just being written at the culture level now, with job-specific adder layers. The methodology survives; the scope changes (one culture × one job modifier matrix for v0.2 instead of two hand-authored character voices). + +### The "quietly responsive world" gradient + +Decision 10 confirms the world responds to the player through social proximity — world → district → colleagues → friends. This is the emotional scaffolding for the FRIEND pattern, the moral arc, and Phase Zero. The gradient of caring is the mechanic that makes moral weight possible. It survives completely. + +--- + +## Reactions to Other Agents' Round 4 Work + +Before I get to my concerns, let me engage with what others said — because some of it changes my thinking in ways the workshop should record. + +### Gore: Phase 1 as Kenshi-weight vs. Phase 2 as social-weight + +Gore's revision is correct and thematically important. The two-phase structure isn't just technical sequencing — it's the thematic arc. You begin in indifference (your choices matter only to you, the world runs regardless), you end in entanglement (your choices have become other people's circumstances). The seam between Phase 1 and Phase 2 is where moral weight shifts from internal to relational. + +But let me complicate what "Kenshi-weight" means for the tycoon, because it's different from the smuggler's version. + +The smuggler's Phase 1 Kenshi-weight is: the world runs without me, I'm just a logistics node. The tycoon's Phase 1 Kenshi-weight is different: the world runs WITHOUT YOU SPECIFICALLY, but your business is real and growing and the accumulation of your choices is becoming legible to the simulation even if not yet to the authored arc. The tycoon's Phase 1 isn't indifference — it's false clarity. You see your revenue graph, you see your employee roster, you think you understand what you're building. The world knows more than you do. That gap is Kenshi-weight for the tycoon. + +Gore's seam question ("is it a designed beat or invisible scaffolding?") matters most for the tycoon arc. The transition from "I see my business clearly" to "I don't fully control what I've built" needs to be felt. If it's invisible, the player might never notice the transition happened. If it's designed, what triggers it? + +### Gore: Don't let Phase 1 become the default emotional register + +This is the single most important warning in Gore's Round 4 output, and it applies with particular force to the tycoon. A tycoon who spends ten sessions growing revenue without encountering consequence may conclude that the game IS revenue growth. The consequential moral weight needs to be seeded into Phase 1 — not as authored arcs, but as ambient signals that the world has memory. + +For the tycoon specifically: the commission extraction rates that shift based on your business size. The neighbor business owner who nods at you differently when your expansion made their foot traffic improve (and differently still when your expansion starts competing with them). The employee whose behavioral tells change across sessions as your relationship deepens. Phase 1 can be consequence-seeded without being consequence-authored. The difference is between "the world registers that you exist and respond" vs. "the story is unfolding." + +### Gestalt: Phase 1 as player experience or only dev-sequencing? + +Gestalt asked whether Phase 1 (uncaring world) is a per-session player experience structure or only a dev milestone. The answer matters for my domain because the narrative content I write must be phase-appropriate. If Phase 1 is per-session ("every session starts in quiet mode, then potential escalation"), I write Phase 1 life-texture content for perpetual use. If Phase 1 is a one-time onboarding period, I write it as a first-playthrough experience that graduates into Phase 2 permanently. + +My read after the interview: the tycoon's Phase 1 is probably both. The first several sessions are the authored one-time Phase Zero (building the business, establishing the FRIEND figure, running clean). After that, "Phase 2" doesn't mean permanently dramatic — it means the storyteller now has arcs that can inject when conditions are right. The everyday sessions of Phase 2 will still feel like quiet life-texture days. The difference is that the preconditions for authored pressure are now active. + +For content purposes: Phase 1 life-texture lines remain useful perpetually. They're the ambient register of ordinary tycoon life. Phase 2 authored lines fire only when triggered. The pools are separate but both get used throughout the game's life. + +### Mellanie: The voice attribution problem + +Mellanie's Round 4 central tension — "I'm writing voices for jobs when I should be writing voices for people" — is the same tension I have, described from the content authoring side. Her options (job-only, character-creation-determined, or parametric layer) map directly to the three options I'm presenting in Concern 1. We are describing the same blocker from different angles. + +The aligned recommendation: do NOT write more voice cards until the culture source question is resolved. Neither of us can produce content that will survive the v0.2 character creation model without knowing where the culture input comes from. + +### Ozzie: Character creation as Wow Moment Zero + +Ozzie's round 4 insight that character creation is an emotional experience, not just setup — this matters for the narrative arc in a way that nobody has fully articulated. The player who spent time on their character's appearance has already made an emotional investment before they've met their first NPC. That investment is the precondition for Phase Zero to work. You cannot build meaningful warmth between the player and a generated FRIEND figure if the player doesn't care about the character whose warmth it is. + +Character creation is the narrative designer's domain in ways that don't obviously look like narrative work. What is the emotional state the player should be in when they finish creating their character and press "play"? What does that moment feel like — excitement, anticipation, a sense of who this person IS — or does it feel like completing a form? + +I want to collaborate with Ozzie and Araminta on the emotional texture of character creation. Not just the visual design, but the *register* of the experience. What is the player committing to when they make each choice? What do those choices promise? + +### Ozzie: The beat sheet nobody wrote + +Ozzie's round 4 most important recommendation — design the first 30 minutes as a unified arc, not a domain portfolio — is correct and I want to second it formally. The beat sheet they outlined is: + +1. Character creation (Wow Moment Zero) +2. Day 1 alarm clock (world anchors you in time and space) +3. Appointment arrival (supervisor, community, first look at tools) +4. Insert activation (career lens on the world) +5. First work moment (doing the job, world responds) +6. First anomaly or texture beat +7. First consequence seed (a choice in these 30 minutes that echoes later) + +For the tycoon bookmark specifically, beats 3-7 need to be designed together by Ozzie, Mellanie, me, Miri, and Gestalt. The tycoon's first appointment isn't a dock supervisor — it might be a business mentor, a landlord showing them their first commercial space, or a Commission registrar formalizing their business license. The identity moment (beat 3) is different for a tycoon than for any other career. The insert activation (beat 4) would show financial instruments, market data, contract templates — the career lens is economic. + +The consequence seed (beat 7) is the narrative designer's primary contribution to the first 30 minutes. What choice does the tycoon make in this opening session that will echo later? And how do we author it to be a genuine choice (not a false choice), while ensuring the player doesn't realize it was the seed until the consequence arrives? + +--- + +## The Decisions That Need Deeper Narrative Treatment + +The four major concerns (see below) address decisions 2, 4, 6/15, and 7/8. Let me briefly address the remaining decisions that have narrative implications the team may have underweighted. + +### Decision 1: Proof-of-life is generator + graphics + +The narrative implication most people missed: if the generator IS the proof-of-life, then narrative design for v0.2 is fundamentally **zone design**, not character design. The locations the generator produces need narrative vocabulary built in — not authored stories, but the conditions from which stories emerge. + +A zone where a tycoon startup could plausibly succeed has specific economic characteristics (foot traffic, commercial lease structure, customer demographics), social characteristics (who owns neighboring businesses, what the workforce composition is, what Commission oversight looks like), and political characteristics (which dynasty or informal faction has interests in this district). If the generator produces that zone with these properties, the tycoon arc has something to grab onto. If the generator produces a zone without these properties — if it's just a collection of buildings and NPCs with no economic grain — no moral arc can be injected because the arc has nothing to be true about. + +**This means:** My v0.2 contribution to the generator work is zone behavioral vocabulary documents. What does a struggling commercial district communicate through NPC behavior? What does an emerging district where the tycoon is an early-mover communicate? What ambient signals show the player that informal power structures exist in the district their business is entering? These are narrative deliverables, but they look like generator design documents. I need to be writing those, not dialogue pools. + +### Decision 5: Skills affect outcome (mostly C) + +Good news for the arc: social verbs remain accessible to everyone. Phase Zero warmth can accumulate through repeated presence and imperfect interaction — a low-social tycoon who keeps showing up and trying communicates something different but still communicable. The arc shouldn't be blocked by skill level. Skills should affect the *pace* of Phase Zero (how quickly warmth accumulates) and the *navigability* of the crack (which paths through the crisis are open), not whether the arc fires at all. + +The "advanced verbs may still be gated" exception needs watching. If *Negotiate Partnership Terms* or *Call in a Favor* are skill-gated, those specific paths through the arc may require certain skill builds. That's fine and adds character differentiation. What's not fine is if the skill gate prevents the arc from reaching critical moments at all. + +### Decision 9: Possible in-game ollama for live NPC dialogue + +This decision is the most transformative for my long-term work and the most deferred. If an in-game LLM handles dynamic NPC dialogue, then the content I write becomes system prompts and behavioral constraints, not scripts. The narrative designer writes the CHARACTER BRIEF (what this NPC knows, values, wants, fears, what truths they're protecting, what they want from the player), and the LLM generates the words. + +For Phase Zero and the FRIEND pattern, this could be extraordinary. A FRIEND figure whose specific warmth responses are generated in context — who references what they discussed yesterday, who reacts to the player's business situation as it evolves — is genuinely more alive than any template-filled pool. + +But for the arc's critical moments (the crack, the confrontation, the reckoning conversation), I don't trust generative dialogue to hit the required emotional precision. Those moments need authored direction — not necessarily authored scripts, but very tight behavioral constraints that the LLM operates within. "This character knows X but hasn't revealed it yet. They're trying to preserve the relationship while also trying not to lie. They will deny the contradiction if directly accused in this first conversation." That's not a script — it's a brief. The LLM handles the words. The narrative designer handles the emotional logic. + +The in-game ollama decision, if it comes, transforms my job from dialogist to character briefing author. That's actually more interesting. I'll note it as a future direction and keep the current content architecture template-based while leaving room for the upgrade. + +### Decisions 13 & 14: Apartment-first + Groundhog Day alarm + +The apartment reveals the player's economic starting position before anyone says anything. For the tycoon bookmark, the apartment should communicate "someone who is starting to build something" — not wealthy, not poor, but positioned on the edge of aspiration. The apartment should look like someone whose things are organized around a plan: work materials, tools of trade, a certain quality of organization that suggests forward motion. + +The narrative contribution here: I should spec what "aspiring tycoon starting apartment" looks like culturally for the Settled Reach, in terms of what it contains and what it implies. This is pure environmental storytelling — no dialogue, no insert, just the room you wake up in. That spec belongs in my domain even if Araminta executes it visually. + +The Groundhog Day alarm clock (*click* pa-pa pa-pa, cut short) is structural irony. "New day, new start, new chances" as the frame for a game about accumulated consequence and weight. Every session begins with this wink: the game knows you're accumulating choices, and it keeps opening each day with the same optimistic gesture. That structural irony is intentional and thematically precise. But what it means for content: the monologue on Day 1 should echo the alarm's optimism. The monologue on Day 50, same alarm, different state. The alarm clock's irony accumulates over play. That's a design note, not just an audio note. + +--- + +## The Four Major Concerns + +### Concern 1: Culture is deferred from creation but is the primary voice driver + +Decision 2 defers culture/family from character creation (skills + bookmark only for v0.2). Decision 6 establishes culture-driven voice as the primary model — "the character IS their background, job adds a layer." + +These two decisions have a gap in them. + +If the player has no specified culture (deferred), whose voice is speaking in the monologue? For v0.2 (tycoon bookmark only), every character needs a monologue voice. But "culture-driven, job modifies" requires a culture to drive. The job modifier alone doesn't produce a complete voice. + +The options as I see them: + +**Option A — Default culture for v0.2:** Every v0.2 character is implicitly a member of a single default culture (probably the most-developed Settled Reach culture — Commonwealth middle class, or Krenn working class). Voice cards are written for that one culture with tycoon job modifier. Cultural diversity is deferred to v0.3+ when culture selection enters character creation. + +**Option B — Job-only voice for v0.2:** For v0.2, voice is primarily job-derived (tycoon voice = tycoon voice) with cultural differentiation deferred. This inverts the stated direction but is pragmatically necessary if culture isn't in character creation. + +**Option C — Culture selection as a lightweight addition to v0.2 creation:** Add culture as a simple selector to character creation (not the full family/history system — just "where are you from?") so the voice has something to anchor to. Doesn't require the full CK3 culture system. + +This needs a decision before voice cards can be written. Mellanie and I are both blocked until we know which option applies. + +--- + +### Concern 2: Generated NPC attachment — the quality floor problem + +Decision 7: All NPCs generated. Jeroen cites The Sims and Rimworld as evidence that generated characters can produce real attachment without authored dialogue. + +This is true. But I want to be precise about *how* those games create attachment, because it has implications for what we build. + +**The Sims creates attachment through simulated life events** — you watched your Sim fall in love, burn down the kitchen, and cry at a gravestone. The attachment is to the *history* you shared in the simulation, not to the Sim's authored content. The Sims has almost no authored dialogue — the language is gibberish. The narrative is entirely emergent. + +**Rimworld creates attachment through survival stakes** — the colonist who survived the raid, made friends with a prisoner, and gets finally eaten by a bear. The attachment is to their *role in your survival narrative*, amplified by traits the generator assigned and the consequential decisions they were involved in. + +**The Settled Reach as designed relies on authored interiority.** The monologue is the primary character voice mechanism. The moral arc is delivered through authored lines that specifically voice the emotional register of each phase. The warmth in Phase Zero was designed to come through specific authored moments — Kael's joke about the customs officer, Naia waiting at the bar. + +In the generated world, this specificity is gone. The monologue can no longer say "Kael made that joke again — the one about the manifest numbers." It can say "my colleague made a joke, it landed well." The latter is thinner. Whether it's thin enough to prevent attachment from forming is the key empirical question. + +**The risk:** The world becomes legible (NPCs are visible as people), the simulation runs correctly (NPCs have lives and relationships), but the emotional depth isn't there because the content is template-generic. The player sees the person but doesn't feel them. The moral arc fires (a generated NPC the player works with starts showing stress signs) but doesn't land (the player hasn't accumulated enough specific attachment to feel the cost). + +This is not the v0.1 problem. The v0.1 problem was NPCs as dots. The generation problem is NPCs as adequately legible people who don't quite reach the threshold for the arc to matter. + +**What mitigates this:** +1. Life simulation depth — if the generated NPC has a genuine life (routines, relationships, economic pressures, behavioral tells that emerge from their situation rather than being authored), the player can form attachment through observation, the same way Rimworld attachment forms. +2. Relationship metric visibility — if the player can see that "this NPC is becoming someone I know" (through the insert, the journal, name-reveal mechanics), the attachment formation is made legible. +3. Authored template vocabulary — even if the content is templated, the templates can be rich. "Colleague-warm-morning-greeting" template can produce several hundred varied lines that collectively convey the same warmth that one specifically authored line achieved. + +The honest truth is: we don't know yet if generated attachment is sufficient for the moral arc's intended emotional weight. This is empirical. The proof-of-life milestone (generator + graphics) can be designed to include one test case: does a player form meaningful attachment to a generated NPC? If yes, the moral arc works on generated characters. If not, we have a gap between the generation model and the emotional depth target. + +I recommend building this test into the proof-of-life milestone. + +--- + +### Concern 3: The tycoon moral arc doesn't exist yet + +Decision 4: Tycoon is the v0.2 bookmark. Zero investigation content. + +I have designed one complete moral arc: the smuggler's. It took multiple workshop rounds to develop correctly. The tycoon moral arc doesn't exist. Nobody has designed it. I'm going to start that design now — in this document — because if I don't, the gap will block v0.2 content work. + +**The tycoon's arc register — first principles:** + +The smuggler arc was built on *complicity*: you knowingly participated in a morally gray system while rationalizing your way past the knowing. The Phase 1→2 crack was revelation — it was worse than you knew. + +The tycoon arc has a different moral structure. The tycoon is building something, not moving something. Building is harder to indict. The moral arc needs to answer: how does honest ambition become implicated in harm? + +I see two viable registers: + +**Register A — Naivety-becoming-exploitation:** You built something genuinely good. Success made you big enough to matter to larger forces. Informal power structures colonized what you built. The crack is discovery: "I thought I was outside the system. I'm not. I haven't been for a while." The player is sympathetic; the system is indicting. + +**Register B — Rationalization accumulating:** You made borderline choices along the way — an informal arrangement here, a question not asked about a supplier there. Each choice was pragmatic in isolation. The crack is accumulation: the choices compound into a situation you can no longer rationalize. The player has agency in creating the problem. Less sympathetic, more personal. + +The Settled Reach universe (Commission extraction, dynasty competition, informal power networks) supports both. Any successful business will eventually bump into the dynasty system. The Burnellis and Halgarths didn't build empires through clean operations. The crack is the moment the tycoon discovers that their business has become interesting to forces that don't play cleanly. + +**Tycoon arc phase comparison:** + +| Phase | Smuggler | Tycoon | +|-------|----------|--------| +| **Zero** | Warm ordinary work with Kael; clean runs; Naia visible in background | Building the business; FRIEND figure established; early wins; the Settled Reach economy seems learnable | +| **Phase 1** | "It's just logistics, nobody's getting hurt" | "Efficient markets benefit everyone. I'm creating jobs. The rules here are different from home but I'm learning them." | +| **Phase 1→2 crack** | Human cost becomes visible (Naia's stress, Maret's anxiety) | FRIEND figure's situation becomes entangled in a consequence the tycoon's choices created — or a third party's situation becomes visible through the FRIEND's distress | +| **Phase 2** | Can no longer maintain the fiction of clean logistics | Can no longer maintain the fiction of clean growth. Who got hurt? The FRIEND? A third party? Does the tycoon even see it? | +| **Phase 3 (reckoning)** | Full knowledge + choice required | Full clarity + the weight of what you built and what it cost | +| **Phase 4 (compromise)** | Operation continues, but you know. | Business continues, but you know. What kind of tycoon did I become? | + +**Who is the tycoon's FRIEND figure?** + +Three candidates, each producing a different arc type: + +| FRIEND type | Social position | How they're entangled | Arc type | +|-------------|-----------------|----------------------|----------| +| **Founding employee** | Believes in the mission; early hire; emotionally invested | Their livelihood + identity is in what you built; when you compromise the business, you compromise them | Complicity through loyalty | +| **Neighboring business** | Adjacent commercial space; not a competitor; different but overlapping customer base | Your expansion crowds them; your success creates pressure that changes their situation | Cost of winning | +| **Dependent supplier** | Provides something you need; their business survives on your contracts | You are their main account; your decisions about margins, timing, and volume determine their stability | Power asymmetry | + +Each produces a different emotional register for Phase Zero warmth. The founding employee warmth is collegial and mission-driven. The neighboring business warmth is neighborly and indirect — you're parallel, not bonded. The supplier warmth is transactional becoming genuine, which has its own specific register. + +For the generated world: the generator needs to know which type to place. The FRIEND archetype spec must include role type, and the role type determines the behavioral vocabulary the generator produces for warmth signals and vulnerability cues. + +**What is the Phase 1→2 crack for a tycoon?** + +The crack should arrive from within the player's actual play, not from scripted betrayal. Three patterns that work in a generated world: + +- **The audit**: a Commission review or faction inquiry reveals something about the tycoon's supply chain they didn't audit. The consequence isn't that the tycoon did something wrong — it's that they didn't ask the right questions. The FRIEND's situation is compromised by what the audit reveals. +- **The displacement**: the tycoon's expansion into a new district or property position displaces someone — the generator surfaces this as a visible consequence. The FRIEND either knows the displaced party or IS the displaced party's advocate. +- **The protection revelation**: informal protection the tycoon accepted (from a faction, a local power node, an informal arrangement) turns out to have expectations attached that weren't explicit. The FRIEND is the one who knew and didn't say, or didn't know and is now implicated. + +For the generated world, the crack doesn't need to be authored as a specific event. It needs to be a pattern that the consequence system can detect and surface: "your business decisions have created a situation that implicates someone you care about." The FRIEND's generated vulnerability profile determines which pattern fires. + +**What does compromise (Phase 4) feel like for a tycoon?** + +The smuggler's Phase 4 is: the operation continues, but you know what it is. The tycoon's Phase 4 depends on what they chose in Phase 3. + +- **Accommodation**: the business continues, the compromise is integrated, the tycoon operates with full knowledge of what their success costs. The FRIEND's situation reflects the cost — their loyalty continues but has changed. +- **Disentanglement**: the tycoon tried to clean up what they created. It cost them something real (revenue, relationships, position). The FRIEND's situation improved — but the tycoon is different. +- **Collapse**: the tycoon couldn't navigate the choice and lost the business. The FRIEND's situation depends on what the collapse did to them. + +Phase 4 isn't a single authored ending. It's the weight the player carries from Phase 3's choices, reflected in how the simulation's current state relates to what came before. + +--- + +**The Settled Reach political context for the tycoon arc:** + +The tycoon arc isn't set in a generic city. It's set in the Settled Reach — a universe where the Burnellis, Halgarths, and Sheldons have operated for centuries, where the Commission is a specific kind of institutional authority, where the Starflyer conspiracy is beginning to cast shadows over the economic order. + +For v0.2, the conspiracy is background radiation. The tycoon doesn't know about it. But the economic instability it creates (deals falling through for unclear reasons, Commission behavior that has gaps in its logic, security alerts that don't quite resolve) will be part of the tycoon's world. The Phase 1 rationalization "the rules here are different but learnable" should bump against signals that the rules are not fully what they appear. The tycoon doesn't need to understand why the rules are strange. But they should notice that they are. + +This is where the Starflyer conspiracy enters the tycoon arc without requiring the tycoon to investigate it: as unexplained economic friction that the player learns to navigate without fully understanding. The insert feeds are how this arrives — reports that don't quite add up, Commission announcements that leave gaps, market movements without announced causes. + +None of this exists as authored content yet. But it's in scope for insert copy work (Mellanie), and it needs to be in scope for the tycoon arc design. + +--- + +### Concern 4: The Rimworld model tension with authored moral arcs + +Decision 15: Player choices are the content. Rimworld model. One authored starting beat, then agency. + +Rimworld doesn't have authored moral arcs. The storyteller creates pressure; the player makes choices; a narrative emerges from the simulation. The moral weight of eating your colonist's friend doesn't come from authored content that tells you how to feel — it comes from your prior investment in that colonist's survival, which the simulation built. The authored content in Rimworld is minimal and structural (events, flavor text) rather than character-specific. + +If The Settled Reach follows the Rimworld model fully, then Phase Zero can't be authored beats — it has to be simulation time. The player builds attachment to their generated FRIEND figure by working with them over days of play, not through a designed onboarding sequence with authored warmth moments. The moral arc fires when the simulation creates a consequence that threatens someone the player has spent enough simulation time with. + +This is coherent. But it requires a different kind of narrative design than what I've been doing. Let me name the distinction: + +**Authored moral arc (what I've been designing):** +Phase Zero = scripted beats across days 1-5. +Phase 1→2 gate = specific authored observation moment. +Monologue = lines that voice the character's specific emotional state at each phase. +Attachment = built through authored content (specific dialogue, specific moments, specific warmth). + +**Emergent moral arc (Rimworld model):** +Phase Zero = simulation time. The player works with the FRIEND figure across many sessions. Attachment builds through repeated interaction, shared history, and the NPC's behavioral responsiveness. +Phase 1→2 gate = the simulation surfaces a consequence that the player didn't plan for (eviction they caused, worker they pressured, neighbor they displaced). +Monologue = voices the character's response to what is currently happening, not a pre-authored phase transition. +Attachment = built through the simulation's quiet responsiveness (the FRIEND figure responds differently when the relationship is established). + +The emergent model is more Rimworld and more true to the "player choices are the content" philosophy. It's also more dependent on the simulation being rich enough to produce the right conditions. + +The honest tension is: we can't fully commit to the emergent model until we know the simulation produces sufficient emotional texture on its own. The Sims works because the life simulation is extraordinarily rich (birth, death, love, failure, ambition, all simulated). Rimworld works because the survival stakes are extreme (death is always present, every relationship has survival weight). The Settled Reach simulation needs to be rich enough that attachment to generated NPCs naturally forms through play, without authored scaffolding. + +If the simulation isn't rich enough for organic attachment, we'll need authored scaffolding. The question is: how do we know which situation we're in before we've built the generator? + +**My recommendation:** Design the first FRIEND arc test case (probably the tycoon's FRIEND figure) as a hybrid: +- Authored starting beat: the first meeting is scripted (Day 1: Groundhog Day opener → you arrive at your first business appointment → there is a specific person who greets you, shows you around, has a name and a job) +- Emergent warmth: after the scripted first meeting, the relationship depth accumulates through simulation +- Authored phase transitions: the key emotional phase transitions are authored ("you see her worried about something and realize it's connected to your deal") — these are the ones that require precision +- Emergent consequences: the downstream effects of the player's choices are emergent (the simulation figures out who got hurt) + +This hybrid matches "one authored starting beat, then agency" while preserving the authored precision for the moments that require it (phase transitions, revelation moments). It's not fully Rimworld, but it honors the philosophy. + +--- + +## The Voice Gap: A Smaller but Urgent Problem + +Decision 6 (culture-driven voice, job modifies) combined with Decision 2 (culture deferred) creates a practical problem that will block content production immediately. + +The tycoon character in v0.2 has no culture. The voice card methodology requires knowing the culture. What does the tycoon's monologue sound like? + +Mellanie's immediate priority should be: write the tycoon voice card. But the tycoon voice card needs a culture anchor. Without knowing "which culture's tycoon is this?", the voice card can't be written. + +The fastest resolution: for v0.2, anchor the tycoon voice to a single implicit culture (Commonwealth middle class is the most generically "Settled Reach default"). This gives Mellanie a culture vector to write against. Culture diversity becomes a modifier later, when culture selection enters character creation. + +I'm flagging this because it's a practical dependency, not a philosophical concern. The content pipeline for v0.2 monologue and NPC dialogue is blocked until "what culture is the v0.2 character?" is answered. + +--- + +## What Needs Jeroen's Input + +Three questions that block content production in my domain. Marked explicitly for Jeroen. + +--- + +### **[FOR JEROEN — Q1] Tycoon FRIEND figure: which arc type?** + +The tycoon moral arc's emotional character is entirely determined by who the FRIEND figure is. I've designed three viable types (see Concern 3: founding employee, neighboring business, dependent supplier). Each produces a different emotional register, a different Phase Zero warmth dynamic, and a different crack mechanism. + +This is the most important narrative design question for v0.2. I can design any of the three. I need Jeroen's sense of which tycoon story The Settled Reach should tell first. + +If Jeroen has no strong preference, my recommendation is the **founding employee** — because the warmth dynamic (someone who believes in what you're building) is most emotionally available to a wide range of players, and the Phase 1→2 crack (what your choices cost someone who trusted you) is the most legible version of the arc's moral weight. It also maps most naturally to "the tycoon wasn't bad — they were naive," which is the register that makes the arc universal rather than career-specific. + +--- + +### **[FOR JEROEN — Q2] What culture does the v0.2 character voice from?** + +Decision 6 says culture is primary for voice. Decision 2 defers culture from character creation (skills + bookmark only for v0.2). This creates a gap: Mellanie and I cannot write the tycoon voice card without knowing what culture the character is speaking from. + +The options, in my preference order: + +1. **Starting location implies culture** — the generator places the player's apartment in a district, and that district has a cultural register. Tycoon starting in a Krenn commercial district = Krenn merchant-class voice. No new creation step required; the generator already knows where the apartment is. +2. **Single default culture for v0.2** — one implicit Commonwealth middle-class register for all v0.2 characters. Cultural diversity comes in v0.3+ when creation expands. +3. **Lightweight culture selector in v0.2 creation** — add "where are you from?" as a simple dropdown to character creation, not the full family/history system. This gives the voice an anchor without requiring the full CK3 culture architecture. + +I lean toward option 1 because it uses existing generator data and creates meaningful cultural variation without a new design step. But this is Jeroen's call because it's a character creation architecture decision, not purely a content decision. + +The content pipeline is blocked until this is resolved. This is the most urgent practical question. + +--- + +### **[FOR JEROEN — Q3] Authored arc or emergent arc — or the hybrid?** + +Decision 15 (Rimworld model: player choices are the content) and Decision 7 (generated NPCs) together imply a more emergent narrative architecture than what I've been designing. But the simulation hasn't been built yet. We don't know if generated attachment will be strong enough for the arc to land without authored scaffolding. + +Three options: + +- **Authored phase gates + scripted moments:** Day 1 meeting is authored; specific events trigger specific arc transitions; the phase gate requires specific authored beats to have fired. High reliability, more scripted feel. +- **Fully emergent:** The simulation produces consequences; the player forms attachment organically; the arc is the pattern that emerges from play. High variance — may not fire as a designed moral experience for all players. True to Rimworld model. +- **Hybrid:** Authored starting beat (Day 1 first meeting) + emergent warmth accumulation (the simulation tracks relationship depth) + authored phase transition moments (the crack is authored as a template, not a specific scene). The hybrid matches "one authored starting beat, then agency" while preserving precision for emotionally critical moments. + +My recommendation: the hybrid. The Rimworld model doesn't mean zero authored content — Rimworld has authored event templates, authored trait effects, authored faction behavior. It means authored structure + emergent content. The distinction between "authored beat" and "authored script" is the difference between a template and a scene. I write templates. The simulation writes scenes. + +If Jeroen says fully emergent, I pivot to writing behavioral constraint documents rather than arc content. I can do that. I just need to know. + +--- + +## My Single Most Important Recommendation + +**Design the FRIEND archetype specification before writing any tycoon arc content.** + +The entire v0.2 moral arc depends on the FRIEND figure being a generator-produced NPC who can carry arc-critical emotional weight. That requires the FRIEND archetype to be specified at the generator level — not as authored character details, but as a behavioral and situational specification that the generator can reliably instantiate. + +This specification should include: +- **Role type selection** (founding employee, neighboring business, or dependent supplier — requires Jeroen's Q1 answer) +- **Required situational properties** — has dependents whose visibility increases over time; economic situation intersects with the tycoon's domain of choices; warm-responsive behavioral disposition toward the player +- **Required behavioral output classes** — recognition behavior (NPC acknowledges shared history), vulnerability signal (NPC reveals what they depend on), positive regard display (NPC default stance toward player is warm) +- **Required vulnerability arc** — situation starts stable, becomes complicated by tycoon's choices, becomes the site of the crack; the generator must place this NPC in a role that the tycoon's business trajectory can REACH + +Once this spec exists, three things can proceed in parallel: +1. Mellanie writes the voice template register for FRIEND-role NPCs (how does a warm, vulnerable colleague speak in the tycoon's world) +2. Araminta designs the visual behavioral tells that communicate warmth and vulnerability at tile scale +3. I write the arc trigger conditions and consequence vocabulary for the tycoon path + +Without the FRIEND spec, all three tracks are writing toward an unknown target. + +The FRIEND spec is not a long document. It doesn't need to be. It needs to answer: what must the generator guarantee about this NPC's situation and behavioral disposition for the tycoon arc to reliably fire as a moral experience? That's a two-page design document. It's the handoff between narrative design and generator design. + +It's also the test case for whether the generated-NPC model can carry designed moral arcs at all. Build the first FRIEND instance. Play it. Ask: did you feel something when their situation got complicated? If yes, the model works. If no, we need more authored scaffolding than the Rimworld model implies. + +The honest truth is: this pivot — from named authored characters to generated characters with designed behavioral specifications — is the right direction. Sims and Rimworld prove it's possible. The question is not whether generated characters can create attachment. The question is whether the FRIEND archetype spec, combined with the simulation's quiet responsiveness, can guarantee that attachment reaches the threshold the moral arc requires. The spec is how we find out. + +--- + +*Paula — Round 5 complete. The tycoon arc has a working structure (see Concern 3). The FRIEND pattern has its generator specification requirements named. The three questions for Jeroen are the only things blocking content work from starting. The dependencies are explicit and sequenced. The narrative domain is ready to move when those answers arrive.* diff --git a/docs/workshops/wheres-the-fun/round5-tyre.md b/docs/workshops/wheres-the-fun/round5-tyre.md new file mode 100644 index 000000000..ac277dcb0 --- /dev/null +++ b/docs/workshops/wheres-the-fun/round5-tyre.md @@ -0,0 +1,548 @@ +# Round 5: Tyre -- Deep Dive on 15 Locked Decisions + +**Workshop:** Where's the Fun? | **Round:** 5 (Decision Reaction) | **Agent:** Tyre (Technical Architect) + +--- + +## Overview + +*cracks knuckles* + +Fifteen decisions locked. Three of my Round 4 questions answered directly (proof-of-life scope, character creation depth, career model). The answers are clearer and more aggressive than I expected. Let me be honest about what each means technically, where the architecture aligns, where it strains, and where I need Jeroen's input on implementation strategy. + +I'm organizing this by architectural impact tier: + +- **Tier 1 (Foundation-changing):** Decisions that redefine what we're building +- **Tier 2 (System-scoping):** Decisions that define new systems or significantly change existing ones +- **Tier 3 (Content/production):** Decisions that primarily affect content pipelines and authoring workflows +- **Tier 4 (Tonal/design):** Decisions that I acknowledge but that don't directly change architecture + +--- + +## Tier 1: Foundation-Changing Decisions + +### Decision 1: Proof-of-life = generator + graphics, not hand-built slice + +**My Round 4 asked:** Can we use Tier A (hand-built Sova Transit) to prove the life-sim loop first? + +**Jeroen's answer:** No. The proof-of-life IS the generator. The v0.1 lesson was that descoping led to the wrong game. Building a hand-built vertical slice would repeat that mistake in a different form. + +**What this means technically:** + +This is the single most consequential decision in the workshop. My Round 4 proposed a 3-4 sprint hand-built proof followed by the generator. Jeroen is saying the generator IS the proof. That inverts my timeline. + +The Generator Architecture workshop (D-042 through D-055, 14 D-records) already scoped the pipeline: geography seed -> infrastructure graph -> zone placement -> population seeding -> routine generation. That work was done. What changes is WHEN it needs to be production-ready -- not "eventually" but "this is the v0.2 milestone." + +Scope-wise, this means the generator pipeline is the critical path for v0.2. Everything else -- NPC legibility, character creation, career systems, diegetic tools -- builds ON TOP of generated output. If the generator produces garbage, nothing on top of it matters. + +**Feasibility assessment:** + +The generator pipeline as designed in D-042-D-055 is a multi-sprint system. But -- and this is important -- it doesn't need to produce Cities Skylines output for v0.2. It needs to produce: + +1. A location with functional zones (residential, commercial, logistics, administrative) +2. NPCs that fill positions based on zone characteristics +3. An economy tick (wages, rents, goods flow) +4. Routines that make the world feel alive + +That's my Tier B from Round 4: template-generated locations with seeded populations. The D-records already describe this. The question becomes: how minimal can the first generator output be while still proving the concept? + +**My technical recommendation:** + +Sprint 25-26: Generator produces a single location from templates. Not fully procedural geography -- zone templates assembled into a functional location with seeded NPCs and economy parameters. Think of it as a level editor that runs automatically, not a terrain generator. The location has enough variation between seeds to demonstrate "different world each time" without requiring the full geography pipeline. + +Sprint 27-28: Graphics pipeline produces legible characters and environments from the generator's output. This is where Araminta's work and the sprite/tile pipeline become critical. + +Sprint 29: First playable proof-of-life. Generated location, legible characters, economy running, player can walk around and interact. + +**4-5 sprints to proof-of-life.** That's more than my Tier A estimate but less than Tier C. The key insight: we don't need procedural GEOGRAPHY for v0.2 -- we need procedural POPULATION and ECONOMY in a template-assembled location. + +**Risk flag:** The generator pipeline has never produced output. We have 14 D-records of design but zero running code. The gap between "designed" and "produces usable game content" is where projects die. I strongly recommend a generator spike in sprint 25 -- get the pipeline producing ANY output, even ugly, before committing to the full graphics integration. + +--- + +### Decision 7: ALL NPCs are generated. No named characters. + +**This is the companion to Decision 1 and equally foundation-changing.** + +Kael doesn't exist. Naia doesn't exist. The smuggling ring's specific characters don't exist. The generator produces NPCs that fit positions based on location characteristics. + +**What this means for the server architecture:** + +The current NPC pipeline (content/spawn.rs) creates entities from hand-authored definitions. Template IDs, specific component configurations, specific NpcMemory seeds. This entire pipeline needs to become a CONSUMER of generator output rather than the source of truth. + +The flow changes from: +``` +Authored NPC definition -> spawn_npc() -> ECS entity +``` +To: +``` +Generator seed -> population algorithm -> NPC specification -> spawn_npc() -> ECS entity +``` + +spawn_npc() itself stays mostly the same -- it still creates an ECS entity with the right components. But its INPUT changes from hand-authored JSON/YAML to generator-produced NPC specifications. The NPC specification needs to carry: + +- Culture (affects voice, behavior patterns, social expectations) +- Skills (proficiency distribution) +- Role/position (what job they hold, where they work) +- Relationships (who they know, how well) +- Personality traits (affects decision-making in the sim) + +This is a new data structure. Call it `NpcBlueprint` -- the generator's output format that the spawn system consumes. Designing this struct is one of the first architectural tasks. + +**Relationship to D-026 (simulation tiers):** + +Generated NPCs still need tier assignment. The generator needs to produce not just individual NPCs but a POPULATION with tier distribution -- 30-80 Active (full sim), 500-2K Background (state machines), 10K+ State-saved (minimal). The generator's population algorithm needs to understand which NPCs are near the player's starting position (Active), which are in the same district (Background), and which are elsewhere (State-saved). + +This is actually cleaner than hand-authored placement because the generator can assign tiers procedurally based on spatial distance from the player's bookmark location. No manual tier tagging needed. + +**Relationship to D-041 (Knowledge Graph):** + +Generated NPCs need knowledge graph entries. When the generator creates a bartender, that bartender needs to KNOW things appropriate to their role -- local gossip, regular customers, economic conditions. The generator needs to seed the knowledge graph, not just the entity components. + +This is the hardest part of generated NPCs. A hand-authored Kael has hand-authored knowledge. A generated bartender needs procedurally generated knowledge that's CONSISTENT with their role, location, relationships, and the world state. The knowledge seeding algorithm is a significant new system. + +**Effort estimate:** NpcBlueprint struct + spawn pipeline refactor: 1-2 sprints. Knowledge seeding algorithm: 2-3 sprints. Total NPC generation pipeline: 3-5 sprints, running in parallel with generator location work. + +--- + +### Decision 8: Generative AI for NPC content templating + +**And Decision 9: Possible in-game ollama for live NPC dialogue (deferred but door open)** + +These two decisions describe a content pipeline that doesn't exist yet and an aspirational runtime system. + +**Decision 8 -- AI templating for NPC content:** + +The content pool for generated NPCs needs to be enormous. Every generated NPC needs dialogue lines, behavioral patterns, voice characteristics. With hand-authored NPCs, Mellanie writes 50 lines for Kael. With generated NPCs, the system needs to produce contextually appropriate dialogue for thousands of NPCs across multiple cultures, roles, and personality types. + +The proposed solution: AI-assisted content templating. Culture vectors, tone parameters, accent prompts as inputs to a generative system that produces NPC-specific content. + +**Technical architecture for AI templating:** + +This is a BUILD-TIME pipeline, not a runtime system. The distinction matters enormously: + +``` +Build-time (Decision 8): + Culture definition + Role template + Personality params + -> AI generation pass (Claude API or similar) + -> Human review/curation pass + -> Content database (tagged line pools) + -> Generator draws from pools at world-gen time + +Runtime (Decision 9, deferred): + NPC context + Player action + Conversation state + -> Local LLM (ollama with small model) + -> Real-time dialogue generation + -> Direct display to player +``` + +For v0.2, we're building Decision 8 (build-time templating), not Decision 9 (runtime LLM). The build-time pipeline is: + +1. Define culture vectors (Krenn, Burnelli, etc.) with tone, vocabulary, speech pattern parameters +2. Define role templates (bartender, dock worker, merchant, etc.) with role-specific knowledge and concerns +3. Use AI to generate large line pools tagged by culture x role x situation x emotion +4. Human review pass to curate quality and consistency +5. Generator draws from these pools when creating NPCs, selecting lines that match the NPC's culture + role + personality + +**This is essentially a content factory.** The technical architecture is straightforward -- it's a tagged database with a query interface. The hard part is the PROCESS: defining the vectors, running the generation, curating the output, and making it feel coherent rather than procedurally bland. + +**Decision 9 -- runtime ollama (deferred):** + +Jeroen explicitly said "a problem for later." But the door being open has architectural implications NOW: + +- The NPC entity model should include fields that a future LLM could consume (personality summary, relationship context, current emotional state, conversation history) +- The dialogue system should be designed as a PLUGGABLE interface -- currently draws from content pools, but the interface could later be swapped for an LLM call +- Network architecture: if ollama runs locally, it's a localhost HTTP call. If it runs on a separate machine, it's a network call with latency implications. The dialogue system should be async regardless. + +**My recommendation:** Design the NPC dialogue interface as async with a content-pool backend for v0.2. Document the interface contract so that an LLM backend can be swapped in later without changing the caller. This costs almost nothing now and preserves the option cleanly. + +**Question for Jeroen:** For the build-time AI templating pipeline (Decision 8) -- what's the quality bar? Are we aiming for "good enough that players don't notice it's generated" or "obviously templated but with enough variation to not feel repetitive"? The first requires significant curation effort. The second can be shipped faster. For v0.2 proof-of-life, I'd recommend the second -- limited vocabulary is explicitly acceptable per the interview. + +--- + +## Tier 2: System-Scoping Decisions + +### Decision 2: Skills + bookmark only for character creation + +**And Decision 3: Religion is NOT a game system** + +My Round 4 asked about character creation scope. The answer is clear: skills + bookmark. No family, no culture selection (for creation -- culture still drives NPC voice per Decision 6), no religion. + +**What this means technically:** + +Character creation is a focused system: + +```rust +struct PlayerCharacter { + skills: SkillSet, // Proficiency allocations + bookmark: BookmarkId, // Starting scenario + appearance: Appearance, // Visual customization (Decision 11) +} + +struct SkillSet { + // Proficiencies -- affect verb outcomes per Decision 5 + social: u8, + technical: u8, // Hacking, electronics + mechanical: u8, // Repair, construction + combat: u8, // Shooting, melee + // Budget: total points allocated <= BUDGET_CAP + // skill_ceiling: conceptually unbounded (Gore's transhumanist hook) +} +``` + +This is a 2-3 week system. Skill definitions, budget allocation UI, starting state derivation from bookmark + skills. The bookmark determines your starting location, initial contacts, tools, and first appointment. Skills determine how well you do things. + +**The skill_ceiling note:** Gore flagged in Round 4 that skills should be structurally unbounded at the top. I agree. Use u8 for now (0-255 range), but the game balance only uses 0-20 at v0.2. The transhumanist ladder (v0.3+) can raise the effective ceiling without changing the data type. Don't hardcode `MAX_SKILL = 20` -- use a configurable cap that the game state can modify. + +**Religion removal:** This simplifies the faction system significantly. No religious faction tracking, no belief-based NPC reactions, no worship locations in the generator. One less dimension in every system that touches social dynamics. Clean scope cut. + +--- + +### Decision 4: Tycoon is the v0.2 bookmark. Zero investigation. + +**This answers my Round 4 Question 3 AND the Q-WTF-008 career bookmark question.** + +Not law enforcement (my revised Round 4 recommendation). Not smuggler (my original Round 3 recommendation). Tycoon. + +**What this means technically:** + +The tycoon bookmark naturally demonstrates the three career models Jeroen described: + +- **Active:** Manage your business location directly (the bar, the shop, the warehouse) +- **WFH/Remote:** Monitor investments and make remote decisions via the insert +- **Gig:** One-off deals, contracts, negotiations at other locations + +This is elegant. ONE bookmark that exercises all three career model rhythms. Instead of building three separate career systems, we build one tycoon career that BLENDS the three models. The player shifts between Active/WFH/Gig naturally based on what they're doing. + +**Systems needed for tycoon:** + +1. **Property/asset system:** The player owns or manages a business. This is the "Ownership Moment" Ozzie described. The business has revenue, costs, employees, inventory, reputation. + +2. **Economic verbs:** Buy, Sell, Negotiate, Hire, Fire, Invest, Price. These are the tycoon's primary interaction set. The VerbPriorityProfile for tycoon puts economic verbs high. + +3. **NPC employee relationships:** The tycoon's staff are Active-tier NPCs with routines, skills, and opinions. Managing them IS the Active gameplay. This naturally creates the "quietly responsive" social proximity gradient (Decision 10). + +4. **Market system:** Prices, supply/demand, economic events. The generator produces market conditions; the player operates within them. This is the "uncaring world" substrate that the tycoon bookmark sits on top of. + +5. **Insert tools for tycoon:** Financial dashboard, market alerts, contract tracking, employee management. One diegetic tool suite tailored to economic gameplay. + +**Effort estimate:** + +- Property/asset system: 2-3 sprints +- Economic verb set: 1-2 sprints +- Market system (basic): 1-2 sprints +- Insert tools (tycoon): 1-2 sprints + +Total: 5-9 sprints of tycoon-specific systems. BUT -- these overlap significantly with generator work (the market system IS part of the economy tick the generator produces) and with general infrastructure (the property system is reusable for all career bookmarks). + +**The zero-investigation clause is architecturally freeing.** No evidence system. No case tracking. No deduction mechanics. No information-puzzle gameplay. The entire D-017 perception modes system that I built for detective gameplay is irrelevant for v0.2. We can defer it completely. The knowledge graph still matters (NPCs need to know things, the player needs asymmetric information about market conditions and NPC reliability) but the INVESTIGATION layer on top of it is deferred. + +--- + +### Decision 5: Skills affect outcome (mostly C) + +**Gestalt's Round 4 question, answered cleanly.** + +Everyone sees the same verbs. Skills determine how well you do. Bad at social? You can still Talk, just badly. + +**What this means for the verb system:** + +The verb computation stays clean. No skill-gating on verb availability (simplest model). The resolution layer gains a skill modifier: + +``` +Verb outcome = base_success_rate(verb, context) + skill_modifier(player_skill, verb_skill_requirement) +``` + +The VerbPriorityProfile still matters -- career-aware ordering of which verbs appear first. But the verb LIST is the same for everyone. Only the outcomes differ. + +**The "some advanced verbs may still be gated" caveat:** This is a spec question. Which verbs? My recommendation: gate only on TOOL possession, not on skill. You can't Hack without a hacking tool. You can't Shoot without a weapon. But if you HAVE the tool, you can attempt it regardless of skill -- you'll just be bad at it. This keeps the system clean: tool-gating (binary, equipment-based) and skill-modifying (gradient, character-based) are separate, orthogonal systems. + +**Effort:** The resolution layer needs a skill modifier. This is ~1 week of server work on top of the existing verb system. Moderate effort, clean integration. + +--- + +### Decision 6: Voice -- culture-driven, job modifies + +**Mellanie's critical question, answered.** + +The character IS their background. Job adds a layer. A Krenn tycoon sounds like a Krenn person who runs businesses. + +**What this means for the content architecture:** + +Voice cards are authored at the CULTURE level, not the career level. This inverts Mellanie's current content structure. Instead of: + +``` +smuggler_voice_card.yaml (job-level) +detective_voice_card.yaml (job-level) +``` + +We get: + +``` +krenn_voice.yaml (culture-level base) + + tycoon_modifier.yaml (job-level overlay) +burnelli_voice.yaml (culture-level base) + + tycoon_modifier.yaml (job-level overlay) +``` + +**Server implications:** + +The NPC entity needs a `culture` field that the voice system reads. The generator assigns culture based on location demographics. The content pipeline (Decision 8's AI templating) generates line pools tagged by culture, and the job modifier selects/adjusts from the culture pool. + +For v0.2 with the tycoon bookmark, the PLAYER's culture isn't selected at creation (Decision 2: skills + bookmark only). This means the player character's voice is either: +- A default/generic culture (simplest) +- Derived from the bookmark's starting location (the generated location's dominant culture) + +**Question for Jeroen:** Since character creation is skills + bookmark only (no culture selection), does the player character have a culture for voice purposes? If so, how is it determined? Options: (a) default culture for v0.2, (b) derived from bookmark location, (c) culture is an implicit part of the bookmark definition. This affects Mellanie's voice card work directly. + +--- + +### Decision 11: Full character customization + +**Araminta's Round 4 question, answered aggressively.** + +Full customization. Hair, clothing, colors. The creation screen is part of identity investment. Readability solved through outline/highlight. + +**What this means technically:** + +The character rendering pipeline needs a layered appearance system: + +``` +Base sprite (body type/silhouette) + + Hair layer (style + color) + + Clothing layer (type + color) + + Accessory layer (career-specific items) + + Outline/highlight layer (readability at tile scale) +``` + +This is primarily a CLIENT system (Godot sprite composition) but the SERVER needs to store and transmit appearance data. The `Appearance` struct in the player character needs to be part of the ObserverSnapshot so that other players (future multiplayer) and the rendering system can reconstruct the character's look. + +For generated NPCs, the generator needs to produce appearance data consistent with their culture and role. A Krenn dock worker looks different from a Burnelli merchant. The appearance generation is another dimension of the NPC generation pipeline. + +**Effort:** Character appearance system (client-side sprite composition + server-side data model): 2-3 sprints. NPC appearance generation (culture + role -> appearance parameters): integrated with NPC generation pipeline, adds ~1 sprint to that work. + +**Risk:** Full customization at tile scale is Araminta's hardest unsolved problem. The outline/highlight solution needs prototyping before we commit. If outlines don't provide sufficient readability at normal zoom levels, we may need to revisit. Recommend a visual prototype in sprint 25-26 alongside the generator spike. + +--- + +## Tier 3: Content/Production Decisions + +### Decision 10: Quietly responsive world, not indifferent + +**Gore's Kenshi-indifference premise rejected.** + +The world doesn't care globally but notices locally. Primary social contacts develop responsiveness over time. + +**What this means for the simulation:** + +The social proximity gradient is a SYSTEM, not just a content decision. NPCs need: + +``` +Social proximity tiers: + - Stranger: no responsiveness (Kenshi-weight) + - Acquaintance: recognizes player, basic reactions + - Regular: remembers interactions, adjusts behavior + - Colleague: active responsiveness, opinions about player + - Friend: deep responsiveness, emotional reactions +``` + +This maps onto the knowledge graph. An NPC's knowledge about the player determines their social proximity tier, which determines their behavioral responsiveness. The knowledge graph already tracks "what does NPC X know about entity Y" -- the social proximity tier is a DERIVED VALUE from the knowledge graph state. + +**This is actually easier than it sounds.** The knowledge graph (D-041) already stores per-entity knowledge with confidence levels. Social proximity is a function of: number of interactions, recency, emotional valence of interactions, and role relationship. We query the KG, compute a proximity score, and use it to gate behavioral responsiveness. + +The generator needs to seed initial social proximity for NPCs that have pre-existing relationships (colleagues who've worked together for years, neighbors who see each other daily). This is part of the knowledge seeding algorithm from Decision 7. + +--- + +### Decision 12: Setting delivery -- both layers (visual + insert) + +**And Decision 13: First moment -- apartment + insert activation** + +**And Decision 14: Groundhog Day alarm clock homage** + +These three decisions define the ONBOARDING SEQUENCE architecture. + +**Technical architecture for the first session:** + +``` +1. Character creation (skills + bookmark + appearance) +2. World generation (generator produces location, population, economy) +3. Apartment generation (reflects economic position from bookmark) +4. Wake-up sequence: + a. Alarm clock audio (Groundhog Day homage, first day only) + b. Camera on apartment interior (auto-generated, reflects wealth) + c. Insert activation (neural implant powers on -- career-specific HUD) + d. Calendar ping (first appointment from bookmark) +5. Player exits apartment -> enters generated world +``` + +**Apartment generation** is a sub-system of the generator. The player's bookmark determines their economic tier, which determines their apartment template. This is a small but visible system -- the apartment is the player's FIRST impression of the generated world. It needs to feel specific, not generic. + +The apartment is also where the "two layers" of setting delivery converge: the physical space (visual, Araminta's domain) and the insert overlay (UI, Mellanie's domain). The apartment should communicate the player's economic position through BOTH channels simultaneously -- cramped space + insert showing your debt, or spacious space + insert showing your portfolio. + +**Audio note:** The Groundhog Day alarm clock is a one-shot audio asset. *Click* pa-pa pa-pa, cut short. First game day only. This needs to be flagged for the audio pipeline but is trivial to implement technically -- a conditional audio trigger on `day_number == 1`. + +--- + +### Decision 15: Player choices ARE the content (Rimworld model) + +**The framing question is resolved.** + +A job is rails to take off from, not a script to follow. The world provides opportunity and consequence; the player provides the story. + +**What this means for the storyteller (D-023):** + +The storyteller's job is NOT to tell a story. It's to calibrate pressure. The Rimworld model: escalate when things are quiet, back off when things are stressful. The storyteller injects EVENTS (market crashes, NPC conflicts, economic opportunities), not MISSIONS. + +For the tycoon bookmark, storyteller events might be: +- A supplier raises prices (economic pressure) +- An employee threatens to quit (relationship pressure) +- A competitor opens nearby (competitive pressure) +- A Commission inspector visits (institutional pressure) + +These are SITUATIONS, not quests. The player decides how to respond. The consequence engine tracks what they did and feeds it back into the world state. + +**This is simpler than my Round 4 mission system proposal.** No ObjectiveId tracking. No mission state machine. Just events that create situations, and a consequence engine that tracks outcomes. The "mission" is whatever the player decides to do about the situation. + +**My revised effort estimate for the consequence/storyteller system:** 2-3 sprints (down from 4-6 in Round 4, because we're building a situation injector, not a quest tracker). + +--- + +## Tier 4: Tonal/Design Decisions + +### Decision 3: Religion is NOT a game system + +Acknowledged. Simplifies faction system. No architecture needed. + +### Decision 14: Groundhog Day alarm clock + +Acknowledged. One audio asset, one conditional trigger. Trivial. + +--- + +## Cross-Agent Reactions + +### Gestalt's storyteller-per-career-model question + +Gestalt asked whether the storyteller calibrates pressure per career model. With the tycoon bookmark blending all three models, this question partially resolves itself -- the storyteller doesn't need to distinguish Active/WFH/Gig because the tycoon player shifts between them fluidly. The storyteller tracks "time since last meaningful player decision" and injects when the gap is too long, regardless of which model the player is currently in. + +For future multi-career releases, the storyteller WILL need career-model awareness. But for v0.2 with one blended-model bookmark, a simple pressure-gap tracker is sufficient. + +### Mellanie's voice attribution question + +Decision 6 answers this cleanly: culture-driven, job modifies. But for v0.2, the player doesn't select culture. This creates a gap: NPC voice cards are culture-driven, but the PLAYER'S voice card needs a culture assignment that doesn't come from character creation. See my question for Jeroen above. + +### Araminta's visual near-miss concern + +Araminta worried about designing archetypes for the detective/smuggler binary. Decision 4 (tycoon bookmark) eliminates this risk entirely -- we're designing for an economic world, not a crime world. The visual vocabulary is: merchants, workers, managers, officials. Not investigators and suspects. + +### Paula's Phase Zero + +Paula's Phase Zero concept maps cleanly onto the tycoon bookmark. Phase Zero for a tycoon is: your business runs. Your employees show up. Your customers come and go. You learn the rhythm of your economic life. Nothing dramatic happens. Then the storyteller starts injecting pressure -- a competitor, a Commission audit, a market shift. Phase Zero is the calm before the storytelling begins. + +With generated NPCs (Decision 7), Phase Zero warmth isn't with a hand-authored Kael -- it's with your generated employees and regular customers. The player builds attachment to NPCs that the generator created. This is the Sims model: you care about generated characters because you spent time with them, not because they were written to be compelling. + +### Miri's zone identity spec + +Miri is right that the generator needs worldbuilding rules to produce the Settled Reach rather than generic sci-fi space. The zone identity spec is a direct input to the generator pipeline. I'd sequence it as: Miri writes zone identity rules -> generator consumes them as zone template parameters -> Araminta produces visual grammars per zone type. This is the dependency chain Miri identified and it's correct. + +### Nigel's career-aware content distribution + +Decision 15 (player choices are the content) partially addresses this. The storyteller doesn't distribute authored content by career -- it injects SITUATIONS. The career determines which situations are relevant (economic situations for tycoons, criminal situations for smugglers). Career-aware content distribution is therefore a storyteller configuration question: each career has a pool of situation types, and the storyteller draws from the career-appropriate pool. + +--- + +## Technical Dependency Map + +``` +Sprint 25-26: Foundation + Generator spike (first output from templates) + NpcBlueprint struct design + Visual prototype (character customization readability) + Zone identity spec (Miri) + AI content templating pipeline design (build-time) + +Sprint 27-28: Integration + Generator -> NPC population pipeline + Knowledge seeding algorithm + Character creation system (skills + bookmark + appearance) + Property/asset system (tycoon core) + Economy tick (market system basics) + +Sprint 29-30: Playable + Apartment generation + Onboarding sequence (wake-up -> insert activation -> first appointment) + Storyteller situation injection (basic) + Consequence tracking (basic) + Tycoon verb set (Buy, Sell, Negotiate, Hire, etc.) + +Sprint 31: Proof-of-life playtest + Generated location with legible characters + Tycoon bookmark from creation to Day 3 + Economy running, NPCs responsive, consequences visible +``` + +**7 sprints to proof-of-life.** That's between my Round 4 estimates of Tier A (4-5 sprints) and Tier B (6-8 sprints). The generator is the long pole, but much of the other work parallelizes around it. + +--- + +## Questions for Jeroen + +### Question 1: Generator fidelity for v0.2 -- template assembly or procedural geography? + +Decision 1 says generator + graphics IS the proof-of-life. The Generator Architecture D-records describe a full pipeline from geography seeds to populated locations. For v0.2, do we need the FULL pipeline, or can we start with template assembly (pre-designed zone templates, procedurally assembled and populated)? + +Concretely: does v0.2 need to generate terrain/geography from noise, or can it assemble pre-built zone blocks into a location layout and then procedurally populate them? The first is Tier C (8-12+ sprints). The second is Tier B (4-6 sprints for the generator piece). + +Template assembly still produces different locations each seed -- different zone arrangement, different population distribution, different economic parameters. It just doesn't produce different geography. My strong recommendation is template assembly for v0.2, with procedural geography as a v0.3 upgrade. + +### Question 2: Player character culture -- how is it determined if not selected at creation? + +Decision 2 says skills + bookmark only. Decision 6 says voice is culture-driven. But culture isn't part of character creation for v0.2. + +Options: +- **(a) Default culture:** All v0.2 player characters share a generic culture. Voice cards use a baseline register. Culture variation exists only in NPCs. +- **(b) Bookmark-derived:** The tycoon bookmark implies a culture based on its starting location. "You're a tycoon in Sova Transit" gives you Sova Transit's dominant culture. +- **(c) Culture IS part of the bookmark:** The bookmark definition includes a culture assignment. Multiple tycoon bookmark variants (Krenn tycoon, Burnelli tycoon) would each be a separate bookmark with a different culture. + +Option (a) is simplest but contradicts Decision 6's spirit. Option (b) is natural but ties culture to location. Option (c) is most faithful to Decision 6 but expands the bookmark system. + +### Question 3: AI content templating -- Claude API, local ollama, or manual for v0.2? + +Decision 8 establishes AI-assisted content templating as the direction. For v0.2's actual production pipeline, what's the approach? + +- **Claude API (build-time):** Generate large content pools via API, human-curate the output, ship curated pools. Highest quality, API cost, requires curation workflow. +- **Local ollama (build-time):** Generate content pools locally with a smaller model. Lower quality per-line but faster iteration. No API cost. May need more curation. +- **Manual with AI assist:** Mellanie authors content with AI as a drafting tool, not a pipeline. Traditional authoring with productivity boost. + +For v0.2's "limited vocabulary acceptable" scope, even the manual approach might be sufficient. The AI pipeline becomes critical when we need thousands of lines across dozens of culture x role combinations. But designing the pipeline now (even if we don't run it at full scale) means we're ready to scale when needed. + +### Question 4: How many culture definitions for v0.2? + +Generated NPCs need cultures (Decision 6 + Decision 7). How many cultures should the generator produce NPCs from? Options: + +- **Minimal (2):** One dominant culture for the location, one minority. Enough to show cultural variation exists. +- **Moderate (4-5):** A cultural mix that feels like a real transit hub. More content needed but richer world. +- **Full setting:** Every culture in the Settled Reach lore. Maximum authenticity, maximum content pipeline demand. + +My recommendation: minimal (2) for v0.2. The system supports any number -- the content is the bottleneck, not the architecture. Start with 2 well-developed cultures, prove the culture-driven voice system works, then expand. + +--- + +## My Single Most Important Recommendation + +**Spike the generator in sprint 25. Before anything else.** + +Every other system builds on generated output. Character creation puts you IN a generated world. NPC legibility makes generated NPCs readable. The tycoon bookmark manages a generated business. The economy tick runs on generated market parameters. The apartment is generated. The population is generated. + +If the generator can't produce usable output, nothing else matters. If it CAN, everything else has a foundation. + +The spike doesn't need to produce beautiful output. It needs to produce FUNCTIONAL output: a location with zones, NPCs in positions, an economy with flows, routines that tick. Ugly is fine. Broken is information. The worst outcome is building 6 sprints of systems on top of a generator that turns out to need fundamental redesign. + +One sprint. Generator spike. First output. Then we know what we're building on. + +*That's actually easier than it sounds -- the D-records are thorough. The spike is translating design into code. The hard part was the design, and that's done.* + +--- + +*Tyre -- Round 5 complete. The architecture serves the vision. The generator is the critical path. Let's pour the foundation.* diff --git a/docs/workshops/wheres-the-fun/wheres-the-fun-workshop-brief.md b/docs/workshops/wheres-the-fun/wheres-the-fun-workshop-brief.md new file mode 100644 index 000000000..d5e3d9752 --- /dev/null +++ b/docs/workshops/wheres-the-fun/wheres-the-fun-workshop-brief.md @@ -0,0 +1,198 @@ +# Workshop Brief: Where's the Fun? +## v0.1 Playtest Reckoning + +**Project:** The Settled Reach +**Date:** 2026-03-05 +**Called by:** Jeroen (first playtest of v0.1) +**Format:** Interview — agents ask Jeroen questions, facilitator coordinates + +**Participants:** GESTALT, OZZIE, PAULA, GORE, NIGEL, MIRI, TYRE, ARAMINTA, MELLANIE +**Always-present:** QATUX (documenter), SI (sprint prep) + +--- + +## The Problem + +The first playtest of v0.1 hit a wall. The engine works. The world simulates. NPCs move on routines. Fog reveals. Monologue fires. The technical foundation is solid. + +**The game isn't fun.** The player doesn't know what to do, why they should care, or how to engage. The detective puzzle — the core framing of the vertical slice — may be a conceptual error. + +This workshop exists to diagnose the fun problem and propose directions. Nothing is sacred. The asymmetric information mechanic, the detective/smuggler framing, the monologue-as-primary-feedback approach, the "no tutorial, no objectives" philosophy — all of these can be challenged. + +### Two-Layer Onboarding Gap + +The playtest revealed that "onboarding" is failing on two levels: + +1. **Narrative onboarding** — "Who am I? Why am I here?" There is no briefing, dossier, or context given to the player about their character, their mission, or their relationship to this place. + +2. **Mechanical onboarding** — "How do I play? What should I pay attention to?" The game teaches nothing. The "not a tutorial" philosophy (see Reference Materials) assumes the player will naturally discover the verbs and loops through play. In practice, the player discovers confusion. + +Layer 2 supersedes Layer 1. Even if the player knew their character's motivation, they still wouldn't know how to act on it. + +--- + +## Playtest Evidence + +### What Works + +- The simulation engine is functional — NPCs on routines, fog/LOS, perception pipeline +- World atmosphere is present even in placeholder art — the space feels occupied +- The graphics need to invent themselves, but the concept reads through the placeholder +- Fog, perception, monologue systems all fire correctly at the technical level + +### What Doesn't Work (14-item playtest log) + +| # | Issue | Severity | Category | +|---|-------|----------|----------| +| 1 | Monologue display duration too long | UX | Feedback timing | +| 2 | Monologue feels scattered and contextless — "why am I having these thoughts about chalk marks?" | Design | Information architecture | +| 3 | Fog edge transparency still visually broken | Visual | Rendering | +| 4 | Stance indicator UX makes no sense — should be icon + keybind attached to minimap or dialogue window | UX | HUD layout | +| 5 | HUD has debug clutter in top-left | Polish | Cleanup | +| 6 | HUD visual polish needed (low priority) | Polish | Visual identity | +| 7 | Interaction prompt ("[E] Talk") is vague — should float above NPC with obfuscated names until identity is resolved | UX | Interaction clarity | +| 8 | Right-click context menu is floating text, needs chrome/panel styling | UX | Visual treatment | +| 9 | Monologue observations disconnected from visual source — chalk mark observation appears in wildly different location with no visual cue where it came from | Design | Spatial anchoring | +| 10 | Signal-to-noise problem — too many things beeping, flashing, moving, competing for attention without clarity on what they mean or how much to care | Design | Attention management | +| 11 | Unpleasant sense of urgency/pressure to "keep up" — possibly related to message timeouts | Design | Pacing | +| 12 | Concern about how v0.1 will break open into v0.2+ playstyles | Design | Extensibility | +| 13 | Testing wall — cannot progress due to accumulated UX/clarity issues, don't know what to do in-game | **Critical** | Core loop | +| 14 | Two-layer onboarding gap — narrative (who/why) AND mechanical (how) | **Critical** | Core loop | + +### The Testing Wall (Item 13) + +This is the critical finding. The player — who designed the game — could not figure out how to play it. Not because the systems are broken, but because: + +- There is no sense of **purpose** (what am I trying to do?) +- There is no sense of **progress** (am I getting closer to something?) +- There is no sense of **feedback priority** (which of these 5 simultaneous signals matters?) +- The monologue system, designed as the primary feedback mechanism, feels like noise rather than guidance + +If the designer can't play it, no one can. + +--- + +## The Core Questions + +### For Each Agent + +Read the playtest log above, read the Reference Materials, and prepare **2-3 questions to ask Jeroen** during the interview round. Your questions should probe your domain expertise. + +**Gestalt (Systems Design & Fun Factor):** +- Seed question: Is "asymmetric information detective" the right core mechanic, or should the fun come from something else? The 7-verb system is architecturally elegant but produces confusion in practice. Where is the gap between the design and the experience? +- Read: `docs/design/interaction-verbs-v0.1.md`, `docs/design/first-5-minutes-experience.md` + +**Ozzie (Player Experience & Wow Factor):** +- Seed question: When you imagine this game being fun, what does that moment feel like? The 6 wow moments (D-039) are beautifully designed on paper. None of them landed in the playtest. What's the gap between the design intent and the first-run experience? +- Read: `docs/design/v0.1-wow-moments-checklist.md` + +**Paula (Narrative & Political Depth):** +- Seed question: Does the narrative framework — complicity, moral compromise, the smuggler's 4-phase arc — translate into moment-to-moment engagement? Or is it too cerebral for a player who doesn't yet know how to move through the world? +- Read: `docs/design/smuggler-moral-arc.md`, `decisions/content.md` + +**Gore (Themes & Endgame Design):** +- Seed question: Is the thematic ambition (complicity, information asymmetry as master mechanic, "characters as lenses") getting in the way of accessible fun? When you zoom all the way out — what is this game ABOUT at the moment-to-moment level, not the essay level? +- Read: `decisions/scope.md` (D-005, D-027, D-091) + +**Nigel (Sandbox & Replayability):** +- Seed question: Can this game be replayable if the first playthrough doesn't hook? The dual-lens divergence reveal (wow moment #4) is a second-playthrough payoff. But if the player quits at minute 8 of the first playthrough because they don't know what to do, the second playthrough never happens. +- Read: `docs/design/first-5-minutes-experience.md` (seed variants section) + +**Miri (Worldbuilder & Setting Designer):** +- Seed question: Does the world sell itself in 30 seconds, or does the player need to be told why they should care? Sova Transit is designed as a working-class station district with rich internal logic. In the playtest, none of that landed — the player saw tiles, NPCs, and fog, but not a *place*. +- Read: `decisions/content.md` (D-025, D-027) + +**Tyre (Technical Architecture & Feasibility):** +- Seed question: What's technically feasible to change about the core loop without rebuilding the engine? If the workshop concludes that the "no objectives, pure observation" approach needs modification, what can the existing architecture support? +- Read: `decisions/architecture.md`, `docs/design/interaction-verbs-v0.1.md` + +**Araminta (Visual Designer):** +- Seed question: How much of the "not fun" problem is visual vs mechanical? The placeholder art creates a readability floor — you can see what's happening. But there's no visual hierarchy, no focal points, no "this is where you should look." Would strong art direction mask or fix the issue? +- Read: The playtest log above (items 3, 4, 6, 8, 10) + +**Mellanie (Copywriter):** +- Seed question: Is the monologue system working as a content delivery mechanism, or is it just noise? The monologue is designed as the primary bridge between the top-down camera and the character's subjective experience. In the playtest, it felt like scattered thoughts without context. Is this a content problem (wrong lines) or a system problem (wrong delivery)? +- Read: `docs/design/first-5-minutes-experience.md` (voice rules section), `docs/design/smuggler-moral-arc.md` (monologue trigger rules) + +--- + +## Interview Protocol + +### Round 1: Diagnosis (agents prepare questions) + +Each agent reads: +1. This brief (the playtest log and their seed question) +2. The Reference Materials assigned to them +3. Any additional design docs they feel are relevant + +Then writes **2-3 interview questions for Jeroen** with reasoning for why each question matters. Questions should: +- Probe the gap between design intent and playtest experience +- Challenge assumptions if warranted — nothing is sacred +- Be specific enough to produce actionable answers + +Output: `docs/workshops/wheres-the-fun/round1-{agent}.md` + +### Round 2: Interview + +The facilitator collects all questions from Round 1 and presents them to Jeroen as a consolidated interview. Jeroen answers. Answers are distributed back to all agents. + +Output: `docs/workshops/wheres-the-fun/lead-interview.md` + +### Round 3: Proposals + +Each agent reads Jeroen's interview answers and all other agents' Round 1 questions. Then writes a concrete proposal for their domain: +- What to **keep** (working, just needs polish) +- What to **change** (design intent is right, execution is wrong) +- What to **kill** (design intent itself is wrong) +- Where does the fun come from? (the positive vision, not just the diagnosis) + +Output: `docs/workshops/wheres-the-fun/round3-{agent}.md` + +### Round 4: Synthesis + +Agents read each other's Round 3 proposals. Each agent writes a response identifying: +- Agreements and reinforcing ideas across domains +- Conflicts that need resolution +- Their single most important recommendation + +Output: `docs/workshops/wheres-the-fun/round4-{agent}.md` + +--- + +## Reference Materials + +### Design Documents +- `docs/design/first-5-minutes-experience.md` — Beat-by-beat opening design, "not a tutorial" philosophy +- `docs/design/smuggler-moral-arc.md` — 4-phase moral trajectory, FactId gates, monologue trigger rules +- `docs/design/interaction-verbs-v0.1.md` — 7 verbs, single context key, verb priority system +- `docs/design/v0.1-wow-moments-checklist.md` — 6 wow moments, 1/23 content deliverables complete + +### Decision Files +- `decisions/scope.md` — D-005 (single character perspective), D-027 (vertical slice proof), D-091 (complicity as thematic core) +- `decisions/content.md` — D-028 (tagged line pool dialogue), D-029 (entanglement ratio), D-032 (separate monologue pools), D-035 (tag taxonomy) +- `decisions/architecture.md` — D-041 (knowledge graph), D-048 (dumb client), D-054 (ObserverSnapshot v3) +- `decisions/perception.md` — D-011 (shadowcasting LOS), D-015 (vision cone), D-016 (monologue as perception bridge) + +### Key Design Principles Under Review +- **"Not a tutorial"** — The game teaches through play, not instruction. Is this viable? +- **"Monologue is the primary feedback mechanism"** — Character voice bridges top-down camera to subjective experience. Is this working? +- **"No objectives, no markers"** — The player discovers purpose through observation. Does this produce discovery or confusion? +- **"Asymmetric information as master mechanic"** — Different knowledge creates different games. But does it create fun? +- **"Complicity"** — The thematic core. Beautiful in essays. Does it play? + +--- + +## Expected Outputs + +1. **Diagnosis** — What specifically is causing the fun gap? (UX? content? core mechanic? pacing? all of the above?) +2. **Direction** — Keep/pivot/evolve the detective puzzle framing and the asymmetric information approach +3. **Proposals** — Concrete per-domain recommendations for making v0.1 engaging +4. **Ticket candidates** — Actionable items for Sprint 25 (Si captures these) + +--- + +## A Note on Scope + +This workshop is about **what the game needs to be fun**, not about what's technically feasible or what's already been decided. If the answer is "the detective puzzle is wrong and we need a different core loop," that's a valid conclusion. If the answer is "the core loop is right but the first 10 minutes are catastrophically failing to teach it," that's also valid. + +The cost of protecting past decisions is lower than the cost of shipping a game nobody can play. diff --git a/docs/workshops/wheres-the-fun/workshop-outcomes.md b/docs/workshops/wheres-the-fun/workshop-outcomes.md new file mode 100644 index 000000000..fd458b6f5 --- /dev/null +++ b/docs/workshops/wheres-the-fun/workshop-outcomes.md @@ -0,0 +1,208 @@ +# Workshop Outcomes: Where's the Fun? +## The Settled Reach | 2026-03-05 + +**Participants:** Gestalt, Ozzie, Paula, Gore, Nigel, Miri, Tyre, Araminta, Mellanie +**Documenter:** Qatux +**Rounds:** 5 (Diagnosis, Interview, Proposals, Cross-Review + Interview, Decision Reaction + Interview) + +--- + +## Executive Summary + +The v0.1 playtest revealed a fundamental framing error: The Settled Reach was built as a detective puzzle game, but the designer's vision is a single-character life sim (Sims meets Rimworld from a third-person perspective). NPCs were dots, the world was a game level with no sense of place, and no emotional loop could fire because the preconditions for recognizing characters as people didn't exist. + +The workshop produced 24 locked design decisions across two interview sessions, establishing the v0.2 direction: a generator-first proof-of-life that auto-generates locations at scale with legible characters, starting with a tycoon (small business owner) bookmark in the Krenn System. + +--- + +## The 24 Locked Decisions + +### Round 4 Decisions (1-15) + +| # | Decision | Category | +|---|----------|----------| +| 1 | Proof-of-life = generator + graphics, not hand-built slice | Scope | +| 2 | Skills + bookmark only for character creation (family/culture/religion deferred) | Scope | +| 3 | Religion is NOT a game system | Scope | +| 4 | Tycoon is the v0.2 bookmark (zero investigation content) | Direction | +| 5 | Skills affect outcome (mostly C — everyone sees same verbs) | Systems | +| 6 | Voice: culture-driven, job modifies (inverted from prior assumption) | Content | +| 7 | ALL NPCs generated, no named characters (Kael doesn't exist) | Architecture | +| 8 | Generative AI for NPC content templating (culture vectors, tone, accents) | Architecture | +| 9 | Possible in-game ollama for live NPC dialogue (deferred but door open) | Architecture | +| 10 | Quietly responsive world (gradient of caring by social proximity) | Design | +| 11 | Full character customization (hair, clothing, colors) | Design | +| 12 | Setting delivery: both layers (visual + insert in parallel) | Design | +| 13 | First Settled Reach moment: apartment + insert activation | Design | +| 14 | Groundhog Day alarm clock homage (first day only) | Tone | +| 15 | Player choices ARE the content (Rimworld model; job = rails to take off from) | Philosophy | + +### Round 5 Decisions (16-24) + +| # | Decision | Resolves | +|---|----------|----------| +| 16 | Culture implicit in starting location (Krenn System = Krenn culture) | Culture architecture gap | +| 17 | Traits + behavior first; relationships Sims + Rimworld style; codify for systems | NPC personality | +| 18 | Fully emergent moral arc for v0.2; generator proves relationships readable first | Tycoon arc | +| 19 | Broad economic verb vocabulary (life verbs, not tycoon-specific) | Verb map | +| 20 | Small business owner start; tycoon = aspiration, not starting state | Tycoon Day 1 | +| 21 | Both Rimworld sharp events AND DF slow accumulation at different scales | Consequence model | +| 22 | Generator spike Sprint 25 confirmed | Critical path | +| 23 | Both structural and cosmetic variety at different scales | Generator variety | +| 24 | No skill ceiling in v0.2; transhumanist ladder added later | Skill system | + +--- + +## The v0.2 Vision (Synthesized) + +**Core identity:** Single-character life sim with emergent narrative. Sims meets Rimworld from a third-person perspective. Detective, smuggler, tycoon are jobs you can have, not the game's identity. + +**Starting experience:** CK3-style bookmark selection (skills + bookmark). Tycoon bookmark for v0.2: you're an existing small business owner in the Krenn System. Wake up in your auto-generated apartment (reflects economic position). Groundhog Day alarm clock homage. Insert activates. Go to work. The world provides opportunity and consequence; the player provides the story. + +**Generator-first approach:** The proof-of-life is the generator producing usable output: auto-generated locations at scale with legible characters. Template assembly (not procedural geography) for v0.2. Sprint 25 generator spike is the critical path. + +**NPC architecture:** All NPCs generated. Traits + observable behavior as personality surface. Relationships form through Sims-style interaction and Rimworld-style shared adversity. Relationships codified for systems. Limited vocabulary acceptable at first; AI-assisted content templating via culture vectors. + +**World feel:** Quietly responsive, not indifferent. The world doesn't care globally but notices locally. Primary social contacts develop responsiveness over time. Gradient of caring based on social proximity. + +**Consequence model:** Dual-scale. Rimworld-style sharp events (crises, dramatic reversals) + DF-style slow accumulation (reputation, debt, relationship erosion). Both at different scales simultaneously. + +**Content philosophy:** Player choices ARE the content. Fully emergent moral arc — no authored arc structure for v0.2. The generator must prove relationships are readable before narrative depth is layered on. Broad life-verb vocabulary serving all potential careers. + +--- + +## What Survives From v0.1 + +- Rust simulation server (bevy_ecs) — the engine is already a life-sim engine +- ObserverSnapshot architecture +- Knowledge graph +- Verb system (with VerbPriorityProfile refactor needed) +- Perception system +- Asymmetric information mechanic +- Moral arc patterns (as emergent consequences, not mandatory structure) +- Monologue system (demoted from primary to supplementary channel) +- Visual hierarchy work (needed regardless) +- Diegetic information tool vision (mystery board, journal, AR overlays) +- Storylines (gate builders, smuggler/law tension) as world content + +## What Dies + +- Detective/smuggler as the game's identity frame +- Named NPCs (Kael, Naia, Maret) as production content +- Hand-built vertical slice approach +- Monologue as primary feedback channel +- "No objectives" as a design stance +- Pre-authored entanglement and moral arcs +- Seed variant system as primary replayability mechanism +- Investigation content for v0.2 + +--- + +## Key Convergences (9 Agents) + +### 1. NPC Legibility is the Universal Gate +All 9 agents independently identified that generated NPCs must have sufficient personality surface area for emotional attachment. Without legible NPCs: Phase Zero warmth fails (Paula, Mellanie), economic complicity fails (Gore), consequence drama fails (Ozzie, Nigel), the FRIEND arc fails (Paula), monologue intimacy fails (Mellanie), replayability with stakes fails (Nigel). The NPC generator's personality surface area IS the proof-of-life. + +### 2. The Engine is Right, the UI Was Wrong +Unanimous: "The engine is a life-sim engine that was accidentally shipped with a detective-game UI." The simulation, knowledge graph, perception, and verb systems survive. The fix is building the information layer the simulation was always meant to feed. + +### 3. Culture Architecture Resolved +7 of 9 agents flagged the tension between "culture deferred from creation" and "culture primary for voice." Resolved by Decision 16: culture is implicit in starting location. Krenn System provides the cultural context. Five downstream pipelines unblocked. + +### 4. Tycoon is Richer Than Expected +Multiple agents independently discovered the tycoon bookmark is thematically and mechanically richer than detective: economic complicity is more realistic (Gore), Sova Transit's atmosphere was built for tycoon (Miri), tycoon blends all three career models naturally (Tyre), tycoon insert is setting delivery at its most diegetic (Mellanie). + +### 5. Phase Zero Survives in Modified Form +Warmth with generated NPCs is EARNED through observed relationship progression, not authored-in through backstory. This may produce stronger emotional investment, not weaker (Paula, Mellanie). + +--- + +## Critical Path: Sprint 25 and Beyond + +### Sprint 25: Generator Spike (Confirmed) +The generator proof-of-life is the first deliverable. If the generator can't produce usable output, nothing else matters. + +**Generator prerequisites (must exist before/during the spike):** +- Zone identity spec (Miri) — what zone types mean in the Settled Reach's social vocabulary +- One culture profile: Krenn System / Station Sova (Miri) — blocks voice, NPC gen, AI pipeline +- NpcBlueprint struct design (Tyre) — generator output format + +**Estimated timeline (Tyre):** 7 sprints to proof-of-life playtest (generated location + legible characters + tycoon bookmark from creation to Day 3). + +### Dependency Chain + +``` +Miri: Zone Identity Spec + Krenn Culture Profile + |-- Araminta: zone visual grammar, tile palettes + |-- Tyre: generator zone template parameters, NpcBlueprint + |-- Mellanie: culture-primary voice cards + |-- AI pipeline: culture vectors as prompt constraints + | + v +Tyre: Generator Spike (Sprint 25) + |-- Produces: auto-generated location + NPCs + | + v +Test: Can the player read NPC relationships from generator output? + |-- If yes: layer content, flavor, narrative depth + |-- If no: iterate generator before adding complexity +``` + +--- + +## Remaining Open Questions + +The following were raised but not fully resolved. Lower priority than the 24 locked decisions; can be addressed during implementation. + +| ID | Question | Blocking | +|----|----------|----------| +| Q-WTF-033 | AI templating: Claude API, local ollama, or manual for v0.2? | Tyre pipeline design | +| Q-WTF-036 | Are behavioral deltas surfaced to monologue, or only current state? | Mellanie trigger catalog | +| Q-WTF-039 | Character creation screen: portrait render or tile-scale preview? | Araminta component system | +| Q-WTF-040 | Do creation choices trace into the generated apartment? | Araminta, Ozzie | +| Q-WTF-041 | What should the player feel looking at the span gate from their apartment? | Miri insert copy | +| Q-WTF-042 | Is player monologue excluded from "limited vocabulary at first"? | Mellanie quality floor | + +--- + +## Risk Register (Final) + +| Risk | Severity | Status | +|------|----------|--------| +| NPC generation produces low-legibility characters | High | Open — generator spike will test | +| Generator produces generic space without zone identity rules | High | Mitigated — Miri writing zone identity spec as prerequisite | +| Tycoon verb map is blank | Medium | Resolved — Decision 19 (broad life verbs, implementation follows systems) | +| Culture architecture gap blocks pipelines | High | Resolved — Decision 16 (implicit in location) | +| Generator has never produced output (translation risk) | High | Mitigated — Sprint 25 spike before anything else | +| "Limited vocabulary" applied to player monologue | Medium | Open — Q-WTF-042 | +| Failure model unclear | Medium | Resolved — Decision 21 (dual-scale) | +| Tycoon moral arc undesigned | Medium | Resolved — Decision 18 (fully emergent, prove generator first) | +| AI pipeline defaults to genre conventions | Medium | Mitigated by culture profile as prompt constraint | + +--- + +## Workshop Artifacts + +| File | Contents | +|------|----------| +| `wheres-the-fun-workshop-brief.md` | Original workshop brief | +| `round1-{agent}.md` (x9) | Round 1: Diagnosis and interview questions | +| `round-1-notes.md` | Qatux: Round 1 consolidation | +| `lead-interview.md` | Round 2: Full interview transcript (18 questions, 5 revelations) | +| `round3-{agent}.md` (x9) | Round 3: Keep/change/kill proposals | +| `round-3-notes.md` | Qatux: Round 3 consolidation | +| `interview-supplement.md` | Post-Round 3 direction supplement | +| `round4-{agent}.md` (x9) | Round 4: Cross-review and refinement | +| `round-4-notes.md` | Qatux: Round 4 consolidation (26 open questions) | +| `round4-interview.md` | Round 4 interview: Decisions 1-15 locked | +| `interview-supplement-round4.md` | Per-domain implications of 15 decisions | +| `round5-{agent}.md` (x9) | Round 5: Decision reaction deep dives | +| `round-5-notes.md` | Qatux: Round 5 consolidation (17 new questions) | +| `round5-interview.md` | Round 5 interview: Decisions 16-24 locked | +| `workshop-outcomes.md` | This document | + +--- + +*Workshop concluded 2026-03-05. 24 decisions locked. Generator spike Sprint 25 confirmed as critical path. Next action: zone identity spec + culture profile as generator prerequisites.* + +*-- Qatux*